Podcast Episode 23

Episode 023 52 min

David M’Raihi: CPSO at Rivian, on Cryptography, Cars, and the Craft of Security

with David M'Raihi · Rivian

0:00 / 52:00

Also on Spotify, Apple Podcasts, YouTube, Podbean, Player FM

Episode 23 of The Security Podcast of Silicon Valley, published October 19, 2022. Host Jon McLachlan talks with David M'Raihi, chief product security officer at Rivian, who was Jon's mentor and gave him his first team to lead. David explains how copy protection and a friend's doctorate pulled him into cryptography, his best day at Symphony in a room full of bankers, and why the step from a technology to a product to a company is where many startups fail. They cover what failure teaches, teams with different strengths, people asking whether their work makes them happy, office days and plant visits, what ten interruptions a day cost, a hard work year in 2001, security for companies with no security team, doing with 30 to 35 people what once took a hundred, public keys since the 70s, the car as a connected device, quantum computers and RSA, homomorphic encryption, and interviews that save 15 minutes for fit.

Episode notes

In this episode of a YSecurity.io production, we dive deep into the world of cryptography and software security with David M'Raihi, the Chief Product Security Officer at Rivian. With a wealth of experience in the field, David shares his insights on the multifaceted challenges and opportunities in protecting innovative technologies in the automotive industry.

David and host Jon McLachlan engage in a fascinating discussion on the essentials of building a robust security team, the intricacies of cryptographic systems, and the potential vulnerabilities that could lead to breaking RSA encryption. Beyond the technicalities, David offers a glimpse into his philosophy on leading an intentional and deliberate life, both within and outside the realm of cybersecurity.

Listeners will gain a rare look into the strategic thinking and technical prowess required to safeguard the future of mobility at Rivian, where cutting-edge vehicles and sustainable technology meet. David's approach to product security encompasses not only the protection of sophisticated software and hardware but also a commitment to fostering a culture of security awareness and resilience within the organization.

This episode is a must-listen for anyone interested in the crossroads of cryptography, automotive innovation, and personal growth. David M'Raihi's journey and reflections provide valuable lessons on leadership, the importance of a proactive security posture, and the pursuit of excellence in one's professional and personal life.

Tune into this YSecurity.io production for an enlightening conversation with David M'Raihi. Whether you're a cybersecurity professional, a technology enthusiast, or someone curious about the intersection of innovation and security in the automotive industry, this episode offers a unique blend of expertise, wisdom, and inspiration. Join us as we explore the complexities of securing the future of transportation and the deliberate choices that shape a meaningful life.

Hosted by Jon McLachlan — co-founder of YSecurity.

The deep dive

From one-time codes to connected cars

David M'Raihi, chief product security officer at Rivian, with Jon on copy protection, a friend's doctorate that made crypto sound fun, the codes on your phone, a room full of bankers at Symphony, the steps from a technology to a company, people asking whether they're happy, office days and plant visits, the cost of ten interruptions a day, a hard year in 2001, security for a company with no security team, public keys since the 70s, the car as a connected device, quantum computers and computing on encrypted data, and interviews that save time for fit.

Episode 23 of The Security Podcast of Silicon Valley aired on October 19, 2022, with Jon McLachlan hosting. The guest was David M’Raihi, chief product security officer at Rivian (Podbean, Apple Podcasts, Spotify). The episode is called “David M’Raihi: CPSO at Rivian, on Cryptography, Cars, and the Craft of Security.” Jon and David had worked together at Apple and again at Symphony, and on the tape Jon calls David his mentor.

Jon spelled out the title, CPSO, for listeners:

“it stands for Chief Product Security Officer, which is a really nice focus.”

Jon McLachlan, at 0:40

This page follows the conversation in order: a section for each part of it, fifteen figures, fifteen clips from the tape, two LinkedIn posts, the full transcript, and five tools: a quiz, a scorecard for security without a security team, a calculator for what interruptions cost, a timeline, and a to-do list. Where the decodes of the tape disagree on a word, the transcript puts its best guess in brackets, and the page doesn’t quote or retell those words.

A chief product security officer

Jon read out David’s path at the start of the show. Director of technology at Gemplus. Vice president of technology at Arxan, a company that obfuscates code. Principal architect at VeriSign. Security architect at Apple. Chief technology officer and vice president of engineering at Verayo. Chief security officer at Symphony. Engineering and technology leader at UnifyID. Head of product security at Pure Storage, right after the company went public, where David started the security team. An advisor to Lacework. And then the CPSO role at Rivian, the electric truck company.

Usually he hears CSO or CISO, Jon said, but CPSO has a nice ring to it.

“Very nice ring to it.”

David M’Raihi, at 0:52

Listen · Jon McLachlan on episode 23, from 0:30
About 25 seconds: Jon on what CPSO stands for. The clip stops on its own. Read the full transcript.

David’s LinkedIn fills in the dates. Apple from April 2010 to June 2011, covering secure implementations on iOS, Mac OS and the hardware platforms, and protocols such as AirPlay, the App Store and FairPlay. Verayo from June 2011 to April 2014. Symphony from October 2014 to May 2017. Pure Storage from February 2018 to January 2021, where the team focused on hardening the software development lifecycle and on certifications such as FIPS and SOC 2. Rivian from January 2021 to December 2023. He also lists Sift Security, as an advisor and investor from October 2014 to July 2018, when Netskope acquired it (LinkedIn). Gemplus merged with Axalto in June 2006 to form Gemalto, a digital security company whose products include smart cards and tokens. Arxan specialized in anti-tamper and digital rights management, and in April 2020 it joined two other companies to form Digital.ai (Wikipedia, Wikipedia).

A timeline. Before 2010: Gemplus, director of technology. Gemplus merged with Axalto in June 2006 to form Gemalto. Before 2010: Arxan, vice president of technology, at a company known for anti-tamper protection. Before 2010: VeriSign, principal architect, and first author of the HOTP standard in December 2005. April 2010: Apple, security architect for iOS, Mac OS and protocols such as AirPlay and FairPlay. June 2011: Verayo, chief technology officer and vice president of engineering. April 2014: Perzo, then Symphony, chief security officer through the deal that created Symphony. May 2017: UnifyID, engineering and technology. February 2018: Pure Storage, head of product security, after the IPO. January 2021, filled: Rivian, chief product security officer, with four product security teams. April 2024: CARIAD, global vice president and lead security, and acting CISO of CARIAD Inc. from April 2026.
From smart cards to cars. Sources: the episode and David's LinkedIn.

The two first met at a security conference. Jon remembered how:

“You actually reviewed one of the papers that I wrote as a graduate student.”

Jon McLachlan, at 2:04

David confirmed that he had.

David earned his PhD at the École normale supérieure, which Jon called the top school in France and compared to Harvard. Wikipedia describes the school as a grande école in Paris and a member of PSL University (Wikipedia). David’s LinkedIn gives the degree as a PhD in mathematics and computer science, with the thesis “Electronic Money and Privacy Regulation” (LinkedIn).

Copy protection came first

Jon asked what first pulled David into security. It started on the other side of the field, David said:

“I was always interested, I would say, by some security aspects”

David M’Raihi, at 4:07

When he was much younger, he did a little hacking. The question that held his interest was simple:

“How do you copy a program that actually is protected?”

David M’Raihi, at 4:27

It was a topic of some interest on the Apple II and Commodore machines. From there he saw that software had ways to protect an application and the intellectual property inside it. Then his interest moved to mathematics and computer science.

Listen · David M'Raihi on episode 23, from 4:07
About 25 seconds: David on copying protected programs. The clip stops on its own. Read the full transcript.

The other side of that question became a business. Arxan, where David was later vice president of technology, built anti-tamper protection for software (Wikipedia).

A friend’s doctorate in cryptography

David was working on a few projects when a friend, preparing a PhD in crypto, started to explain what crypto was about.

“And I was like, wow, that sounds kind of very cool, right?”

David M’Raihi, at 5:13

It had applications and it had theory. What hooked him was one sentence in French in his friend’s PhD work. David translated it as crypto is fun.

Listen · David M'Raihi on episode 23, from 5:01
About 40 seconds: David on the friend who got him into crypto. The clip stops on its own. Read the full transcript.

Back then, he said, crypto always meant algorithms and protecting data:

“there was nothing like cryptocurrencies or anything.”

David M’Raihi, at 6:03

Things changed a lot in the 10 or 15 years before the tape. He also explained a second word. Cryptology combines building systems with attacking systems, which is called cryptanalysis.

Three cards side by side. Cryptography: building systems, the algorithms that protect data. Cryptanalysis, dashed: attacking those systems, to find where they break. Cryptology, filled: both together, as David described it. Below, a list. When David started, crypto meant algorithms and protecting data. Cryptocurrencies came later, and the word changed a lot in the 10 or 15 years before the tape. Filled: a friend's PhD work got him hooked, and he finished his own PhD at the École normale supérieure.
Building and breaking, and the word for both. Source: the episode.

David got hooked and finished his own PhD. His LinkedIn names his advisor, Jacques Stern (LinkedIn). Wikipedia describes Stern as a cryptographer and a professor at the École normale supérieure, known for the cryptanalysis of many encryption and signature schemes, and he received the CNRS Gold medal in 2006 (Wikipedia).

“that’s how I went into security, more by the, I would say, the research angle.”

David M’Raihi, at 6:35

The codes on your phone

The episode doesn’t get to it, but David’s name is on two standards that many people use every day without knowing it.

In December 2005, the IETF published RFC 4226, “HOTP: An HMAC-Based One-Time Password Algorithm.” David is its first author, listed with VeriSign, and his coauthors were M. Bellare, F. Hoornaert, D. Naccache of Gemplus, and O. Ranen (RFC Editor). In May 2011 came RFC 6238, “TOTP: Time-Based One-Time Password Algorithm,” again with David as first author (RFC Editor). A third, RFC 6287, published in June 2011, describes OCRA, the OATH challenge-response algorithm, with David again among the authors (IETF).

Here is how they work. In HOTP, the server and the device share a secret key and a counter. Both compute a code from the key and the counter, and each side moves its counter forward on its own. Each counter value is used once, which makes each code good for one use (Wikipedia). TOTP replaces the counter with a value based on the current time. Its defaults are six digits, the SHA-1 algorithm, and a new code every 30 seconds. Authenticator apps set it up by scanning a QR code (Wikipedia). Both belong to the work of OATH, the Initiative for Open Authentication, an industry-wide collaboration on open standards for strong authentication (Wikipedia).

Two rows of boxes joined by arrows. HOTP, RFC 4226, December 2005: secret key, shared by both sides. Counter, moves after each use. HMAC-SHA-1, filled, then truncated. A code, good for one use. TOTP, RFC 6238, May 2011: secret key, set up by a QR code. The clock, a step every 30 seconds. HMAC-SHA-1, filled, the default. A code, six digits by default. Below, a list. Both sides compute the code from the key and the counter, and each moves its counter on its own. TOTP replaces the counter with a value based on the current time. Filled: David is first author of both standards, which belong to the work of OATH, the Initiative for Open Authentication.
A shared key, a counter or the clock, and six digits. Sources: RFC 4226, RFC 6238 and Wikipedia.

On April 24, 2024, David shared both RFCs on LinkedIn as a blast from the past. He wrote that most of the authors didn’t expect the algorithms to still be in use for strong authentication 10, 15, even 20 years later, counting from when HOTP was first shared in 2004. He has a running joke with friends that he is “(partly) responsible for those 4 and 6-digit codes.”

David on the RFCs behind one-time passwords.

David M'Raihi · April 24, 2024
David's post on HOTP and TOTP. View on LinkedIn.

A room full of bankers

Jon asked for David’s best day in the valley. David couldn’t pick a single day, but a couple stood out. One came at Symphony, when a small startup called Perzo became part of something bigger, and:

“that day when we were sitting in that big room with all the bankers”

David M’Raihi, at 10:33

He has had a day like that maybe two or three times in his life, he said.

Listen · David M'Raihi on episode 23, from 10:13
About 35 seconds: David on his best day. The clip stops on its own. Read the full transcript.

Jon asked him to tell the story. Perzo focused on one part of communication:

“secure communication, efficient and private between not just two parties, but multiple parties.”

David M’Raihi, at 11:14

A text message is one thing, David said. Then come attachments, and then voice and video. The founders had many ideas, and the pivot was to financial applications:

“because there is really a need for security and privacy and also compliance.”

David M’Raihi, at 11:49

That is hard, he said, and not many companies wanted to take it on. Jon was there too. They sold to firms like Goldman Sachs, BlackRock and JPMorgan, he said.

Wikipedia describes Symphony as an instant messaging and communication service built for financial firms, with chat, voice, video and encryption. Its technology started inside Goldman Sachs. In October 2014, Goldman Sachs and 14 other financial institutions invested $66 million to set up Symphony Communication Services and acquire Perzo, which David Gurle had founded in 2012 (Wikipedia). David’s LinkedIn lists him as chief security officer at Perzo from April 2014 and at Symphony from October 2014, covering product security, from the design of the crypto libraries to code review, plus corporate and infrastructure security (LinkedIn).

Four boxes joined by arrows. Two parties, a private message. Many parties, secure group talk. Attachments, files in the thread. Voice and video, calls, kept private. Below, three cards. Who needed it, filled: financial firms, for security, privacy and compliance. October 2014: Goldman Sachs and 14 other institutions invest $66 million, set up Symphony and acquire Perzo. David's role: chief security officer at Perzo from April 2014, then at Symphony to May 2017.
From two people texting to a room of banks. Sources: the episode, Wikipedia and David's LinkedIn.

From a technology to a company

The step from Perzo to Symphony led David to a point about startups:

“So I think moving from just having the technology to having a product and then having a company, that’s what makes a difference.”

David M’Raihi, at 12:00

That step is hard, he said, and it’s why many companies fail. Failing teaches a lot, he said, more than succeeding. Jon agreed. When you succeed:

“it just goes right to your head”

Jon McLachlan, at 12:31

David went on. When you succeed fast, things happen and there’s no time to digest them. Sometimes it’s luck, or the stars align, or 20 years of work meet the right idea at the right time. When you fail, you pause and try to understand what went wrong. There is usually a lesson, you grow, and the next challenge gets easier. Jon added that this takes a growth mindset, and some people don’t have it.

Listen · David M'Raihi on episode 23, from 12:00
About 30 seconds: David on technology, product and company. The clip stops on its own. Read the full transcript.

Then David came back to the three steps. There is a lot of confusion there, he said:

“people think that they have a company, and actually they just have a technology.”

David M’Raihi, at 14:25

Or they have a product with no market. A real company comes from a set of problems you have identified, and a solution better than the ones that exist, or the first one. Without that angle, you can build something, but it won’t be solid. Jon said that’s:

“perhaps the difference between a company and a hobby.”

Jon McLachlan, at 15:12

There is nothing wrong with a great idea that others use, David said. The open source movement lets people offer their technology, and many companies were built on it. Some people are good at business, some are researchers, and some are talented engineers who, asked to go and market their work, would say they have:

“zero interest in doing any of that.”

David M’Raihi, at 16:20

Jon called it what makes teams stronger than the sum of their parts. David added that the diversity that matters most is in how people approach things. That comes from each person’s experience and background, he said, and it brings a lot to any team.

Three boxes joined by arrows. A technology, not yet a product. A product, needs a market. A company, filled, solves a set of problems. Below, a list. Dashed: first stall, people think they have a company, and they have a technology. Dashed: second stall, a product with no market. Filled: what works, problems you have identified, and a solution better than the ones that exist, or the first one. Failure teaches more than fast success: you pause, find the lesson and grow. Not everyone wants to build a company. Business people, researchers and engineers make a team stronger together.
Three steps, and where companies stall. Source: the episode.

The question people started asking

Jon asked whether David looks for diverse backgrounds when he builds a team. David’s first problem is simpler:

“It’s so difficult to find quality people.”

David M’Raihi, at 17:41

The two years before the tape had brought so many shocks, he said, from the economy to events in Europe to COVID, that nobody could predict the next year. People reacted by asking basic questions they had not had the time, or the wish, to ask before. He saw it even more in younger people, who have more flexibility. Why do this job, or why work with this person? People accepted a lot more a few years earlier, he said, and now they don’t. The power had swung from one extreme to the other, and he expected it to rebalance a little. But the question will stay at the back of people’s minds:

“am I really happy? Do I want to do that?”

David M’Raihi, at 20:10

Listen · David M'Raihi on episode 23, from 19:58
About 15 seconds: David on the question that stays. The clip stops on its own. Read the full transcript.

Jon tied it to a life with purpose. He said it matters that we live:

“lives that are deliberate and intentional and aligned with our own personal values.”

Jon McLachlan, at 20:16

For some people, that means solving a hard open problem. For others, it means working to provide for the people they love. David agreed that there is nothing wrong with that.

Work itself had changed too, David said. A couple of years before the tape, choosing your days at home wasn’t an option, and Jon added that you’d have been an oddball for asking. Some jobs can’t be done remotely, David said, and many can, by people with the mindset for it. Work is just more open now, he said.

Three cards side by side. Before: people accepted a lot more from a job. After two years of shocks, filled: people ask whether the job and the people around them are right for them. Next, dashed: some rebalancing, he expected, and the question stays. Below, a list. The first problem is finding quality people at all. Younger people, with more flexibility, ask the question more. Choosing your days at home wasn't an option a couple of years before the tape.
What changed for the people you want to hire. Source: the episode.

Office days and plant visits

Jon asked whether Rivian had a flexible work policy, and whether David worked from home. David tries to go into the office every week. He asked himself how many days he needed there, and found value in meeting his teams. Still:

“I think I can perform the vast majority of what I do remotely.”

David M’Raihi, at 23:34

The office isn’t the only place that counts. David had visited the plant a couple of times, and he feels he should go at least twice a year, to see and understand how the vehicles get made.

Listen · David M'Raihi on episode 23, from 23:33
About 35 seconds: David on visiting the plant. The clip stops on its own. Read the full transcript.

Other companies are different, he said. If you specialize in cloud security and build on AWS or GCP, the office is mostly about connecting with people. That connection matters at two moments. In a creative phase, sitting together at a whiteboard is easier. Jon said it’s faster and you get to the point sooner, and David added that you can see how people react. In a crisis, a call with a bunch of people on Zoom works:

“but it’s a little dry.”

David M’Raihi, at 25:28

“being in the room can help.”

David M’Raihi, at 25:34

In an execution phase, when there are few surprises and people know what to do, the office helps less. There, people get interrupted, and some have a hard time switching context.

Three cards side by side. A creative phase, filled: together at a whiteboard, faster, and you see how people react. A crisis, filled: a Zoom call works but feels dry, and being in the room helps. An execution phase, dashed: few surprises, so the office mostly brings interruptions. Below, a list. David went into the office every week, and could do most of his work remotely. He visited the plant, and aimed for at least twice a year, to understand how the vehicles get made. For a team that builds in the cloud, the office is mostly about connecting with people.
When the room helps, and when it gets in the way. Source: the episode.

The cost of a context switch

Jon knew the problem firsthand. He had announced publicly that he switches context slowly:

“I can do one thing very well”

Jon McLachlan, at 26:12

He needs focus, he said, and asked people to make sure an interruption was important. David remembered, and described what that means for a person like Jon:

“being interrupted like 10 times during the day”

David M’Raihi, at 26:30

That will hurt the person’s productivity significantly, he said.

Listen · Jon McLachlan and David M'Raihi on episode 23, from 26:05
About 35 seconds: Jon on slow context switching, and David on ten interruptions a day. The clip stops on its own. Read the full transcript.

The calculator below puts a number on it for your team, starting from David’s ten interruptions a day.

Four cards side by side. 10 a day: interruptions per person, from David's example. 10 minutes: to get back into the work, a placeholder. 1.7 hours, filled: of focus each person loses a day. 3,067 hours, filled: a year, for a team of 8 over 230 working days. Below, a list. At $100 an engineering hour, that comes to $306,667 a year, or 76.7 engineer-weeks. Dashed: Jon called himself a slow context switcher, and asked people to interrupt only for important things.
Ten interruptions a day, added up over a year. Sources: the episode and the calculator's placeholder defaults.

The hardest day, and the network after it

Jon asked about David’s worst day in Silicon Valley. The real answer was personal, David said:

“The worst days were always because of personal things”

David M’Raihi, at 29:06

“personal stuff is what matters the most.”

David M’Raihi, at 29:31

For a work day, he picked one in 2001. As the market melted down, the company he worked with cut back, like many others, and closed its Bay Area office. For a couple of months he had no job and didn’t know what came next.

Jon asked whether he wanted to write his own story or to see how this one ended. A network worked differently then, David said. There was no LinkedIn and no social networks, just a phone, a few people on email, and drinks. Still:

“I was not scared or worried.”

David M’Raihi, at 32:58

“It was more like, okay, what are the next steps, right?”

David M’Raihi, at 33:00

He knew a few people, reconnected with them, and joined a project that was starting. That was his rebound, and another project followed. Once you’re connected and have experience, he said, it all comes together.

Listen · David M'Raihi on episode 23, from 32:58
About 35 seconds: David on rebuilding after 2001. The clip stops on its own. Read the full transcript.

Then Jon said something to his listeners:

“David here has actually been my mentor through my career.”

Jon McLachlan, at 34:27

From David, Jon learned leadership styles, security technologies and how to work with customers. And:

“you gave me my first opportunity to really lead a team”

Jon McLachlan, at 34:42

David said Jon was too kind:

“I learned a few things from you as well”

David M’Raihi, at 35:05

Security before the security team

Jon asked David to look into the future and describe a security industry he’d be proud of. David named one trend he sees as positive, with many companies working on it: SecDevOps, or security DevOps. It’s a broad term, he said, but he likes the space, because security has plenty of room to optimize. What has been missing is how to spend resources:

“When do we need a tool? When do we need a human?”

David M’Raihi, at 37:26

In vulnerability management, there are many players. You can run scans, do pen testing and analyze the code. Lacework came up, and David said it does more than that, in cloud security and in automating detection. The angle David cares about is a company like this:

“I don’t have a security team.”

David M’Raihi, at 38:16

It is starting to build products and asks what to do. Not the bare minimum, David said, but:

“what would be a good solution for me, knowing that I can’t hire 50 people?”

David M’Raihi, at 38:24

Such a company may not even be able to hire one security person, Jon said.

Listen · David M'Raihi on episode 23, from 38:05
About 25 seconds: David on a company with no security team. The clip stops on its own. Read the full transcript.

Today the field calls this DevSecOps. Wikipedia describes it as bringing DevOps into security practices: each delivery team builds the right security controls into its own delivery, and tests earlier in the lifecycle, which is called shifting left. Testing covers three areas: static, software composition and dynamic (Wikipedia). The scorecard below asks how much of that a small company has in place.

Fewer people, more tools

In security, people with experience and the right skills are hard to find, David said, and hard to afford. For a long time:

“we are kind of brute-forcing everything.”

David M’Raihi, at 39:03

You need 20 analysts, a security operations center, and more. Now there are options to be smarter. You still need a team, he said, but he can do today with 30 or 35 people what took about a hundred 20 years earlier.

Listen · David M'Raihi on episode 23, from 39:15
About 25 seconds: David on doing more with fewer people. The clip stops on its own. Read the full transcript.

Jon asked whether software-as-a-service security tools explained it. It’s a combination, David said: skills, experience and knowledge, and people who can do several things. People used to be more specialized. Now he sees engineers who handle infrastructure, operations and testing.

Two cards side by side. Twenty years earlier, dashed: about a hundred people. Brute force: 20 analysts, a security operations center, and more. At the time of the tape, filled: 30 to 35 people, working smarter, with engineers who handle infrastructure, operations and testing. Below, a list. Decide what a tool should do and what needs a person. A small company may have no security team, and may not be able to hire even one person for it. Vulnerability management has many players: scans, pen tests and code analysis.
A hundred people then, 30 to 35 now. Source: the episode.

From public keys to email

Security isn’t a brand new industry, David said. Take encryption:

“the first time you had some public encryption algorithm, we are talking, you know, the 70s.”

David M’Raihi, at 40:37

The idea was a public key to encrypt and a private key to decrypt. Then people saw that the same keys could sign. It started with a few people in the mid-70s, he said, and the military and governments were a different story.

Then came the 80s, and email in the 90s. Once people could communicate, he said:

“then there are security issues associated with that.”

David M’Raihi, at 41:39

And security companies grew up alongside. Wikipedia gives the dates. Whitfield Diffie and Martin Hellman published the idea of public-key cryptography in 1976, and introduced digital signatures. Ron Rivest, Adi Shamir and Leonard Adleman publicly described RSA in 1977. An equivalent system was developed in secret at GCHQ, the British signals intelligence agency, by Clifford Cocks in 1973, and declassified in 1997 (Wikipedia). Taher Elgamal described ElGamal encryption, built on the Diffie-Hellman key exchange, in 1985 (Wikipedia). Taher was the guest on episode 61.

Two rows of boxes joined by arrows. Encryption: a message, in the clear. Public key, filled, anyone can encrypt. Ciphertext, safe to send. Private key, filled, only the owner opens. Signatures: private key, filled, the owner signs. A signature, sent with the message. Public key, filled, anyone can verify. Below, a timeline. 1973, dashed: GCHQ, Clifford Cocks develops an equivalent of RSA in secret, declassified in 1997. 1976: Diffie and Hellman publish the idea of public-key cryptography, and digital signatures. 1977, filled: RSA, Rivest, Shamir and Adleman describe it publicly. 1985: ElGamal, Taher Elgamal describes his encryption system, built on the Diffie-Hellman key exchange.
One key to lock, one to open, and the dates. Sources: the episode and Wikipedia.

Cars as connected devices

People should be humble about all this, David said, when they hear that something doesn’t work yet. Take cars. They have been around for more than a hundred years, and an electric vehicle is built differently, he said:

“it’s more like a connected object”

David M’Raihi, at 42:15

“that’s a shift again.”

David M’Raihi, at 42:20

People need to understand what it takes to make that work, he said.

Listen · David M'Raihi on episode 23, from 41:56
About 35 seconds: David on the car as a connected object. The clip stops on its own. Read the full transcript.

Wikipedia describes Rivian as an American electric vehicle maker and automotive technology company founded in 2009. Deliveries of its R1T pickup began in October 2021, and on November 10, 2021, it became a public company through an IPO (Wikipedia). David’s LinkedIn describes the product cybersecurity group he led there in four teams. One handled security integration and validation in the factory and the plant ramp-up. One did offensive security: pen tests of the vehicles and their connected infrastructure, an internal red team, and threat analysis and risk assessments. One did embedded work: hardening the vehicle’s electronic control units, cryptography and hardware security modules. And one built the infrastructure, on premises and in AWS and GCP, for key management and for secure communication between vehicle and cloud, charging and service (LinkedIn).

The rules changed around the same time. Wikipedia’s article on automotive security notes that in June 2020, a United Nations forum on vehicle regulations released R155 and R156, with performance and audit requirements for car makers on cybersecurity and software updates (Wikipedia).

Four boxes joined by arrows. The vehicle, filled, control units and keys. The cloud, vehicle to cloud. Charging, and service. The factory, and the plant ramp-up. Below, four cards. Integration and validation: security in the factory and the plant ramp-up. Offensive security: pen tests of vehicles and infrastructure, a red team, threat analysis and risk assessments. Embedded: hardened control units, cryptography and hardware security modules. Infrastructure: key management on premises and in AWS and GCP, and secure links from vehicle to cloud. Below, a list. Filled: cars are more than a hundred years old, David said, and an electric vehicle is built more like a connected device. In June 2020, a UN forum released R155 and R156, rules for car makers on cybersecurity and software updates.
The car as a connected device, and the four teams that secured it. Sources: the episode, David's LinkedIn and Wikipedia.

When the hard problems get easier

Jon asked whether they would see RSA broken in their lifetimes. David pointed to quantum computers and post-quantum algorithms:

“certain things are going to be more vulnerable.”

David M’Raihi, at 43:00

It wasn’t clear to him yet whether those attacks really work, he said. Jon checked that he meant quantum computing as the thing that could break RSA. David answered that a lot of recent work went into algorithms based on different problems. The traditional ones come from number theory. RSA rests on factoring: there’s no way to extract the two numbers whose product you see. Much of internet security was built on two such problems, Jon added. So people asked what happens in a world where:

“these problems are not as hard as we think”

David M’Raihi, at 44:22

He didn’t think anyone needed to worry about it yet. He left open whether that changes in 20 years. New algorithms exist, he said, and there will be more.

Wikipedia fills in the threat and the answer. Peter Shor’s 1994 quantum algorithm finds the prime factors of an integer, and Shor also gave an algorithm for the discrete logarithm problem. Beating classical computers may take quantum computers with millions of qubits (Wikipedia). On August 13, 2024, NIST released its first three post-quantum standards. FIPS 203, ML-KEM, is the primary standard for general encryption. FIPS 204, ML-DSA, is the primary standard for digital signatures. FIPS 205, SLH-DSA, is a backup in case ML-DSA proves vulnerable. On March 11, 2025, NIST picked HQC, a code-based algorithm, as a backup for ML-KEM, with a draft standard expected in early 2026 (Wikipedia).

Three cards side by side. Factoring: RSA rests on it, find the two primes behind a large number. Discrete logarithms: another hard problem that Shor's algorithm also solves. Shor's algorithm, 1994, filled: a quantum algorithm for both. It may take millions of qubits to beat classical computers. Below, NIST's replacements, four compartments. ML-KEM, filled, FIPS 203, encryption. ML-DSA, filled, FIPS 204, signatures. SLH-DSA, FIPS 205, backup. HQC, dashed, backup for ML-KEM. Below, a list. NIST released the first three standards on August 13, 2024, and picked HQC on March 11, 2025. On the tape, David didn't think anyone needed to worry yet, and left open what 20 years would bring.
The problems keys rest on, the algorithm that breaks them, and the replacements. Sources: the episode and Wikipedia.

Computing on encrypted data

David is more interested in what new primitives make possible. Homomorphic encryption is his example. You produce encrypted objects, and:

“I can operate on these encrypted objects without having to decrypt them.”

David M’Raihi, at 45:15

“It’s almost fascinating”

David M’Raihi, at 45:25

It keeps everything secure along the way, he said.

Listen · David M'Raihi on episode 23, from 44:59
About 35 seconds: David on homomorphic encryption. The clip stops on its own. Read the full transcript.

Jon said that’s what Zama is working toward, and brought up Benoît, whom they knew from Apple. Benoît Chevallier-Mames, then team lead of machine learning at Zama, was the guest on episode 20, on open source homomorphic encryption. Wikipedia credits Craig Gentry with the first plausible construction of a fully homomorphic encryption scheme, in 2009, using lattice-based cryptography (Wikipedia).

Four boxes joined by arrows. Your data, in the clear. Encrypt, filled, with your key. Compute, on encrypted objects. Decrypt, filled, only the result. Below, a list. The data stays encrypted while it is used. Craig Gentry described the first plausible fully homomorphic scheme in 2009, using lattice-based cryptography. Filled: Zama works on it. Benoît Chevallier-Mames, whom Jon and David knew from Apple, was the guest on episode 20.
Compute on the ciphertext, decrypt only the answer. Sources: the episode and Wikipedia.

The script, then the person

Jon asked what David looks for when he interviews, and for his favorite question. Most companies have a process, David said, to be fair and to compare candidates for a role. There’s a set of questions and a panel of different people. That’s hard to do in a fair and balanced way, and most companies do it reasonably well.

He goes a little off script. In a 45-minute interview, he spends 30 minutes on the questions his talent team needs answered. He keeps the other 15 minutes, or 20 in an hour, for personal questions about how the person thinks. What he’s after is fit with the company, the team and the organization’s values.

When a hire didn’t work out, it was rarely about skills, he said, unless you hire a sailor to bake a cake. If someone is good in C and wants to learn, the gap is:

“something that you can easily fix over six months or a year.”

David M’Raihi, at 48:29

A big gap in values is different. Jon put it as a good person who isn’t the best fit. When David hesitates about a candidate, it’s rarely about technical skill. Most of the time, 80 or 90 percent, it’s a gut feeling about fit, he said.

“And it goes both ways”

David M’Raihi, at 49:24

Hiring the wrong fit does the person a disservice. They try their best, and it doesn’t work. That’s why he’s careful, he said.

Listen · David M'Raihi on episode 23, from 48:59
About 40 seconds: David on fit. The clip stops on its own. Read the full transcript.
A 45-minute interview in two compartments. 30 minutes: the talent team's questions. 15 minutes, filled: how the person thinks. Below, three cards. Skills: a gap in skills can be fixed in six months to a year. Fit, filled: when he hesitates, 80 to 90 percent of the time it's about fit with the team and its values. Both ways, dashed: a wrong fit is a disservice to the person, who tries their best.
Thirty minutes for the panel's questions, fifteen for the person. Source: the episode.

Pick a direction early

Jon asked David for words of wisdom he’d give his younger self.

“As soon as you can, find your North Star, and don’t listen to people if they tell you it’s not the North.”

David M’Raihi, at 50:18

“Just ignore them and go after it.”

David M’Raihi, at 50:26

Jon told listeners not to hesitate, and to be deliberate and intentional. A big mistake is better than doing nothing, David said, and Jon added that you only regret the action you don’t take. David agreed, and gave an example:

“it’s always the books that you decide not to buy that you regret, never the books that you bought.”

David M’Raihi, at 50:59

Listen · David M'Raihi and Jon McLachlan on episode 23, from 50:17
About 25 seconds: David's parting advice, and Jon's. The clip stops on its own. Read the full transcript.

What happened after the tape

From Rivian to CARIAD. David’s LinkedIn lists him as Rivian’s CPSO until December 2023. In April 2024 he joined CARIAD, Inc. as global vice president and lead security. He started as chief product security officer for its software-defined vehicle hub and later served as interim vice president of software development for the hub. His main focus, he writes, is to “integrate AI technologies to improve our SDLC,” with an interest in AI agents for vulnerability management, incident response and risk scoring. From April 1, 2026, he is also the acting CISO for CARIAD Inc. (LinkedIn). Wikipedia describes Cariad as the Volkswagen Group’s automotive software subsidiary. It was set up in 2020 as the Car.Software Organisation, took the name Cariad in March 2021, from “Car, I Am Digital,” and develops software for the group’s cars, from driver assistance and infotainment to vehicle controls and online services (Wikipedia).

Building a team again. On April 24, 2024, the same day as his post on the one-time password RFCs, David wrote that building a security organization is probably one of his favorite projects at work, and that he was hiring again across product security: coding, testing, validation, offensive security and integration. He added: “Bonus points if you drive a cool car.”

David on building a security organization again, and hiring.

David M'Raihi · April 24, 2024
David's post on building a security team again. View on LinkedIn.

Volkswagen invests in Rivian. On June 25, 2024, the Volkswagen Group announced that it would invest up to $5 billion in Rivian. The two then launched a joint venture to develop next-generation software-defined vehicle platforms, combining Rivian’s software with Volkswagen’s scale and manufacturing (Wikipedia). That links the company David left with the group behind the one he joined.

Lacework joins Fortinet. On the tape, Lacework wasn’t public yet. In June 2024, Fortinet acquired Lacework, “a data-driven cloud security company,” for an undisclosed amount (Wikipedia). David’s LinkedIn lists him as a Lacework advisor from June 2019 to May 2022.

Rivian’s next model. Rivian started deliveries of its R2, a midsize SUV, in June 2026 (Wikipedia).

More on this site. On episode 61, Taher Elgamal, the cryptographer behind ElGamal encryption, explained why SSL required a certificate from the server and only offered one to the client, which is how the web ended up with passwords. On episode 92, Jasson Casey of Beyond Identity argued that the trouble with passwords, API keys and session cookies is that they move. On episode 25, Jon traced the path of David Gurle, who founded Perzo, from Microsoft and Skype to Symphony and Hive. And on episode 20, Benoît Chevallier-Mames of Zama talked about open source homomorphic encryption.

A timeline. April 2024, filled: CARIAD, David joins as global vice president and lead security. April 24, 2024: two posts, David shares the one-time password RFCs, and says he's building a security team again. June 2024: Lacework, Fortinet acquires the company David advised. June 25, 2024: Volkswagen announces it will invest up to $5 billion in Rivian, then starts a joint venture with it. August 13, 2024: post-quantum, NIST releases its first three standards. March 11, 2025: HQC, NIST picks a code-based backup for ML-KEM. April 1, 2026, filled: acting CISO, David adds the role at CARIAD Inc. June 2026: the R2, Rivian starts deliveries of its midsize SUV.
After the tape, from CARIAD to the R2. Sources: David's LinkedIn and Wikipedia.

Did you catch it? A seven-question quiz

Seven questions from the episode. Each answer links the moment in the transcript or its section.

Quiz · seven questions

Did you catch it?

Pick an answer to see whether you caught it.

  1. What did David want to know when he was much younger?

    a. "How do you copy a program that actually is protected?" Transcript, 4:27.

  2. How did David translate the line in French from his friend's PhD work?

    b. He translated the line as crypto is fun. See A friend's doctorate in cryptography.

  3. Why did financial firms need what Perzo and Symphony built?

    c. "because there is really a need for security and privacy and also compliance." Transcript, 11:49.

  4. Where does a real company start, according to David?

    a. A technology alone isn't a company, and a product needs a market. See From a technology to a company.

  5. How many interruptions a day did David use as his example?

    b. "being interrupted like 10 times during the day" Transcript, 26:30.

  6. How many people does David need today for work that took about a hundred 20 years earlier?

    c. With 30 to 35 people, he could do what took about a hundred 20 years earlier. See Fewer people, more tools.

  7. In a 45-minute interview, how many minutes does David keep for personal questions?

    b. The other 30 go to the questions his talent team needs answered. See The script, then the person.

Can you run security without a security team?

Eight questions for a company like the one David described: building products, with no security team yet. Each tip links the section behind it.

Scorecard · eight questions

Can you run security without a security team?

Answer for your company as it runs today. If two answers fit, pick the lower one.

  1. Does one person own security by name, with time set aside for it?

    • Nobody owns it
    • Shared, informally
    • A named owner, with time on the calendar

    David's question starts with a company that can't hire a team. See Security before the security team.

  2. Do admin, cloud and code accounts all require a second factor?

    • Passwords only
    • Some accounts
    • All of them, with one-time codes or passkeys

    The six-digit codes come from standards David helped write. See The codes on your phone.

  3. Do code and dependency scans run on every build?

    • No scans
    • On some repositories
    • Every repository, every build

    Static and software composition testing are two of the three areas DevSecOps covers. See Security before the security team.

  4. Does a tool watch your cloud configuration?

    • No
    • Someone checks by hand now and then
    • All the time, with alerts someone reads

    Cloud security and automated detection came up with Lacework. See Security before the security team.

  5. Has someone outside tested your product in the last year?

    • Never
    • More than a year ago
    • In the last year, with the findings fixed

    Pen testing is part of the vulnerability work David described. See Security before the security team.

  6. Does every vulnerability get an owner and a deadline?

    • No
    • The serious ones
    • All of them, tracked to closure

    When do you need a tool, and when a human? See Security before the security team.

  7. Do your engineers get hours without interruptions?

    • Rarely
    • Some days
    • Most days, with set times for questions

    Ten interruptions a day add up. See The cost of a context switch.

  8. Do you know where your products rely on RSA or other public-key algorithms that quantum computers threaten?

    • No
    • Partly
    • Yes, with a plan for the new NIST standards

    The first post-quantum standards came out in 2024. See When the hard problems get easier.

Waiting for a team

Security waits for a hire that may not come soon. Name an owner, and turn on a second factor for every admin account this week.

Partly covered

Some of the basics run on their own, and some depend on someone remembering. Next, put scans in every build and give every finding an owner and a deadline.

Covered without a big team

Tools do the repeat work, and people handle the rest. Keep the outside test yearly, and start planning for the post-quantum standards.

What do interruptions cost your team?

David pictured someone interrupted ten times a day. This calculator adds up the time it takes your team to get back into the work after each one.

Calculator

What do interruptions cost your team?

Five numbers: the people on the team, the interruptions each gets a day, the minutes it takes to get back into the work, what an engineering hour costs you, and working days a year.

$306,667a year in lost focus
3,067 hof focus lost a year
76.7engineer-weeks, at 40 hours a week
1.7hours each person loses a day

Interruptions cost your team 3,067 h of focus a year, or 76.7 engineer-weeks, at a cost of $306,667. That's 1.7 hours a day for each person.

The calculator counts the time to get back into the work, and leaves out the interruptions themselves and the mistakes they cause. The defaults are placeholders: ten interruptions a day comes from David's example on the tape, and ten minutes to get back into the work is a guess. Put in your own numbers.

From public keys to post-quantum

The cryptography David talked about, his roles, the companies around him, and this show. Filter by topic.

Timeline

From public keys to post-quantum

  1. A secret first

    Clifford Cocks develops an RSA-equivalent system at GCHQ, declassified in 1997 (Wikipedia, on this page).

  2. Public keys

    Whitfield Diffie and Martin Hellman publish the idea, and digital signatures (Wikipedia).

  3. RSA

    Rivest, Shamir and Adleman describe it publicly (Wikipedia).

  4. ElGamal

    Taher Elgamal describes his encryption system (Wikipedia).

  5. Shor's algorithm

    A quantum algorithm for factoring, and for discrete logarithms (Wikipedia, on this page).

  6. HOTP

    RFC 4226, with David as first author (RFC Editor, on this page).

  7. Gemalto

    Gemplus and Axalto merge (Wikipedia).

  8. Rivian

    The company is founded (Wikipedia).

  9. Fully homomorphic encryption

    Craig Gentry's first plausible construction (Wikipedia, on this page).

  10. Apple

    Security architect, to June 2011 (LinkedIn).

  11. TOTP

    RFC 6238, the time-based version, again with David as first author (RFC Editor).

  12. Verayo

    CTO and VP of engineering, to April 2014 (LinkedIn).

  13. Perzo

    Chief security officer, until Symphony bought it (LinkedIn).

  14. Symphony

    Goldman Sachs and 14 other institutions invest $66 million and acquire Perzo, and David becomes Symphony's chief security officer (Wikipedia, on this page).

  15. Pure Storage

    Head of product security, to January 2021 (LinkedIn).

  16. R155 and R156

    UN rules for car makers on cybersecurity and software updates (Wikipedia).

  17. Rivian

    Chief product security officer, to December 2023 (LinkedIn, on this page).

  18. The R1T

    Rivian starts delivering its pickup to customers (Wikipedia).

  19. Rivian's IPO

    Rivian becomes a public company (Wikipedia).

  20. Episode 20

    Benoît Chevallier-Mames of Zama (episode page).

  21. Episode 23

    David on the show (Podbean, transcript).

  22. Episode 25

    David Gurle, who founded Perzo (episode page).

  23. CARIAD

    Global vice president and lead security (LinkedIn, on this page).

  24. A blast from the past

    David shares the HOTP and TOTP RFCs (LinkedIn).

  25. Fortinet buys Lacework

    For an undisclosed amount (Wikipedia).

  26. Volkswagen and Rivian

    Volkswagen announces it will invest up to $5 billion (Wikipedia).

  27. Post-quantum standards

    NIST releases FIPS 203, 204 and 205 (Wikipedia).

  28. Episode 61

    Taher Elgamal on SSL and passwords (episode page).

  29. HQC

    NIST picks a code-based backup for ML-KEM (Wikipedia).

  30. Acting CISO

    For CARIAD Inc., alongside his role (LinkedIn).

  31. Episode 92

    Jasson Casey on secrets that move (episode page).

  32. The R2

    Rivian starts deliveries of its midsize SUV (Wikipedia).

What to do this week

Five things, in the order of the episode.

  1. Turn on a second factor everywhere. Put one-time codes or passkeys on every admin, cloud and code account, as in The codes on your phone.
  2. Protect time for focus. Set times for questions, so nobody on the team gets interrupted ten times a day, as in The cost of a context switch.
  3. Name an owner for security. Even before you can hire a team, one person should own it, with tools doing the repeat work, as in Security before the security team.
  4. List where you rely on RSA. Find the products and services that depend on public-key algorithms, and plan for the NIST post-quantum standards, as in When the hard problems get easier.
  5. Get help with it. Our Product Security work secures what you build, Penetration Testing tests it from the outside, and vCISO Services give you a security leader before you can hire one. If you want to talk it through, book 15 minutes with us.

Key takeaways

  1. A technology isn't a company

    Moving from a technology to a product, and then to a company, is the hard step, David said. Many startups mistake a technology or a product with no market for a company.

  2. Failure teaches more

    Fast success leaves no time to digest what happened. When you fail, you pause, look for the lesson and grow, and the next challenge gets easier.

  3. Teams need different strengths

    Business people, researchers and engineers each bring something different. The diversity that matters most is in how people approach problems.

  4. The room matters at certain moments

    Being together helps in a creative phase and in a crisis. In an execution phase, the office mostly brings interruptions.

  5. Interruptions cost focus

    Ten interruptions a day hurt a person's productivity in a significant way, David said, and Jon calls himself a slow context switcher.

  6. Security can start before the security team

    A company building its first products may not be able to hire even one security person. Decide what a tool should do and what needs a person.

  7. Do more with fewer people

    David could do with 30 to 35 people what took about a hundred 20 years earlier, with better options and engineers who can do several things.

  8. Hire for fit, train the skills

    A gap in skills can be fixed in six months to a year. When David hesitates about a candidate, 80 to 90 percent of the time it's about fit.

Chapters

Tap a chapter to play from there.

Pull quotes

“it stands for Chief Product Security Officer, which is a really nice focus.”

Jon McLachlan ·

“How do you copy a program that actually is protected?”

David M'Raihi ·

“that day when we were sitting in that big room with all the bankers”

David M'Raihi ·

“So I think moving from just having the technology to having a product and then having a company, that's what makes a difference.”

David M'Raihi ·

“it just goes right to your head”

Jon McLachlan ·

“am I really happy? Do I want to do that?”

David M'Raihi ·

“I can do one thing very well”

Jon McLachlan ·

“being interrupted like 10 times during the day”

David M'Raihi ·

“personal stuff is what matters the most.”

David M'Raihi ·

“When do we need a tool? When do we need a human?”

David M'Raihi ·

“we are kind of brute-forcing everything.”

David M'Raihi ·

“it's more like a connected object”

David M'Raihi ·

“I can operate on these encrypted objects without having to decrypt them.”

David M'Raihi ·

“I don't have a security team.”

David M'Raihi ·

“As soon as you can, find your North Star, and don't listen to people if they tell you it's not the North.”

David M'Raihi ·

About the guest

David M'Raihi, Chief Product Security Officer, Rivian

David M'Raihi was chief product security officer at Rivian when this episode aired, from January 2021 to December 2023. He holds a PhD in mathematics and computer science from the École normale supérieure and is first author of RFC 4226 and RFC 6238, the HOTP and TOTP standards behind many one-time codes. He worked at Gemplus, Arxan, VeriSign, Apple, Verayo, Perzo, Symphony, UnifyID and Pure Storage. Since April 2024 he has been global vice president and lead security at CARIAD, Inc., and from April 1, 2026 also its acting CISO.

About the host

Jon McLachlan is co-founder of YSecurity, the on-demand cybersecurity team for startups, and of Cyberbase.ai. The Security Podcast of Silicon Valley is hosted by Jon McLachlan and Sasha Sinkevich. Jon McLachlan on LinkedIn ↗

Context and sources

  • Episode 23 on Podbean: David M'Raihi: CPSO at Rivian, on Cryptography, Cars, and the Craft of Security. Source ↗
  • David M'Raihi's LinkedIn experience: CARIAD from April 2024, Rivian from January 2021 to December 2023, Pure Storage, UnifyID, Symphony, Perzo, Verayo, Apple, and a PhD from the École normale supérieure. Source ↗
  • RFC 4226, December 2005: HOTP, an HMAC-based one-time password algorithm, by D. M'Raihi, M. Bellare, F. Hoornaert, D. Naccache and O. Ranen. Source ↗
  • RFC 6238, May 2011: TOTP, a time-based one-time password algorithm, by D. M'Raihi, S. Machani, M. Pei and J. Rydell. Source ↗
  • RFC 6287, June 2011: OCRA, the OATH challenge-response algorithm. Source ↗
  • David M'Raihi on LinkedIn, April 24, 2024: a blast from the past, on the HOTP and TOTP RFCs. Source ↗
  • David M'Raihi on LinkedIn, April 24, 2024: building a security organization again, and hiring across product security. Source ↗
  • Wikipedia, HMAC-based one-time password: a shared secret key and a counter, each counter value used once. Source ↗
  • Wikipedia, Time-based one-time password: HOTP with the counter replaced by the current time, six digits, SHA-1 and 30 seconds by default. Source ↗
  • Wikipedia, Initiative for Open Authentication: an industry-wide collaboration on open standards for strong authentication. Source ↗
  • Wikipedia, Rivian: founded in 2009, R1T deliveries from October 2021, the IPO on November 10, 2021, Volkswagen's investment of up to $5 billion announced on June 25, 2024, and R2 deliveries from June 2026. Source ↗
  • Wikipedia, Cariad: the Volkswagen Group's automotive software subsidiary, named Cariad in March 2021. Source ↗
  • Wikipedia, Symphony Communication: a messaging service for financial firms, set up in October 2014 with $66 million from Goldman Sachs and 14 other institutions, which acquired Perzo. Source ↗
  • Wikipedia, Gemalto: formed in June 2006 by the merger of Axalto and Gemplus. Source ↗
  • Wikipedia, Arxan Technologies: anti-tamper and digital rights management, part of Digital.ai from April 2020. Source ↗
  • Wikipedia, Jacques Stern: a cryptographer and professor at the École normale supérieure, CNRS Gold medal in 2006. Source ↗
  • Wikipedia, École normale supérieure: a grande école in Paris and a member of PSL University. Source ↗
  • Wikipedia, Fortinet: the acquisition of Lacework in June 2024. Source ↗
  • Wikipedia, RSA cryptosystem: described publicly in 1977, an equivalent developed at GCHQ in 1973, and Diffie and Hellman's 1976 idea of public-key cryptography. Source ↗
  • Wikipedia, ElGamal encryption: described by Taher Elgamal in 1985. Source ↗
  • Wikipedia, Diffie-Hellman key exchange: published in 1976. Source ↗
  • Wikipedia, Shor's algorithm: a 1994 quantum algorithm for factoring and discrete logarithms. Source ↗
  • Wikipedia, NIST Post-Quantum Cryptography Standardization: FIPS 203, 204 and 205 released on August 13, 2024, and HQC picked on March 11, 2025. Source ↗
  • Wikipedia, Homomorphic encryption: computation on encrypted data, and Craig Gentry's 2009 construction. Source ↗
  • Wikipedia, Automotive security: UNECE regulations R155 and R156, released in June 2020. Source ↗
  • Wikipedia, DevOps: DevSecOps, with security testing in static, software composition and dynamic areas. Source ↗

Transcript

Automatic transcription. Jon McLachlan hosts this episode with David M'Raihi, chief product security officer at Rivian. There is no diarization on the tape, so speaker labels were assigned from the content and checked against reference recordings of each voice from this episode, and a few short lines may sit with the wrong speaker. Names and terms were normalized where the recognizer misheard them (Jon McLachlan, David M'Raihi, Rivian, Ecole normale superieure, Gemplus, Arxan, VeriSign, Verayo, Symphony, Perzo, UnifyID, Lacework, Jacques Stern, ElGamal, Zama). Words in [brackets] are best guesses where the audio is unclear. Answers are transcribed as spoken, with most repeated words and false starts kept. Short backchannels such as "yeah" and "right" are left out where they overlap the other speaker.

Read the full transcript 7,179 words

Jon McLachlan

Hello, everyone, and welcome to another episode of the Security Podcast in Silicon Valley. I'm here today with a very special guest, David M'Raihi.

David M'Raihi

Wow, that's nice.

Jon McLachlan

Yeah. Welcome, welcome to the show, David.

David M'Raihi

Thank you. I mean, many times, [David, you] just butchered my name, but no, no.

Jon McLachlan

Did I get it right? David M'Raihi?

David M'Raihi

Yeah, yeah, you got it right.

Jon McLachlan

I hope so. We've worked together enough. I better be able to [say so]. David is the CPSO at Rivian, the electric truck company. CPSO, for those of you who don't know, it stands for Chief Product Security Officer, which is a really nice focus. Usually I just hear CSO or CISO, but CPSO has a nice ring to it, doesn't it?

David M'Raihi

Very nice ring to it. I like it.

Jon McLachlan

David actually received his PhD from École Normale Supérieure. Did I say that right?

David M'Raihi

Yeah, no, yeah, no, no, no, no, no. It's, yeah, no, it's fine.

Jon McLachlan

I think, I think most people would [unclear] this beautiful French school, which is, by the way, the top school in France.

David M'Raihi

[A] graduate school.

Jon McLachlan

It's extremely good. It's extremely, it's the top. It's the best. It's like the French version of Harvard, right?

David M'Raihi

Yeah, only better. But yeah, yes, of course. Yeah, exactly.

Jon McLachlan

Okay. You were a director of technology at Gemplus. You were the VP of technology at Arxan, the obfuscation company. You were the principal architect at VeriSign. You were a security architect at Apple. This is where we first met. How amazing.

David M'Raihi

Actually, we first met at that security conference, didn't we?

Jon McLachlan

Yes, I didn't, I didn't really. You actually reviewed one of the papers that I wrote as a graduate student. Thought it wasn't so bad.

David M'Raihi

I did review that paper.

Jon McLachlan

It's a small world, isn't it? You were the CTO and VP of engineering at Verayo. You were the chief security officer at Symphony Communications, a secure, seamless communication company. We worked together there again, when we sold all of our software to the Goldman Sachs, the BlackRock, the JP Morgans [of] the world. You're the engineering and technology leader at UnifyID. You're an advisor and investor at [Sift] Security, the head of product security at Pure Storage, right after they had gone public, started their security team. An advisor at [Colsci], which actually sounds like a really cool company. An advisor at Lacework, a very respectable up-and-coming security company. Are they public yet? Lacework?

David M'Raihi

No, not yet. Not yet.

Jon McLachlan

Okay. I know it. I know that you would know.

David M'Raihi

So, and now everybody would know [it's public], [it's public] information, right?

Jon McLachlan

Everyone will know. That's right. But I think you're keeping a special close eye on it, right? So, and then now the CPSO at Rivian. Tons of awesome experience, all in the security world.

David M'Raihi

Yeah, 100% security, that's for sure.

Jon McLachlan

It's a great pleasure to have you on the show, David. Thank you so much for joining us.

David M'Raihi

Thank you, Jon. Really an honor.

Jon McLachlan

So, if you wouldn't share with our listeners, what was the initial nudge into security? How did you get sucked into this game? Maybe something [unclear]?

David M'Raihi

Well, actually, it was funny, the way it happened. I was always interested, I would say, by some security aspects, but more like the other side. When I was way, way, way younger, I was, I would say, hacking a little bit, you know. There were things that I was curious about. How do you copy a program that actually is protected? So there was something that was of some interest, you know, on Apple II and Commodore and [Amiga] machines. And then, you know, from there it was, okay, well, you seem to have some way of protecting an application, protecting some IP. But then I was more interested in kind of mathematics and computer science. What happened is that I was working on a few projects, and there is this friend who was actually preparing his PhD in crypto. And so he starts, you know, explaining a little bit what crypto is about. And I was like, wow, that sounds kind of very cool, right? I mean, it's kind of in between [math] and computer science, and there are some applications, and there is some theory. But the way, really, I mean, it got me is that, you know, on his PhD [thesis], he had this one sentence, [La crypto, c'est rigolo], which in French means, you know, crypto is fun. [unclear], what is it? Crypto is fun. It's fun.

Jon McLachlan

Crypto is fun. Yeah.

David M'Raihi

Yeah. So cryptography is fun. And at the time, obviously, I mean, cryptography and cryptology, which is kind of the combination of building systems and attacking, cryptanalysis, and [them]. And at the time, I mean, there was nothing like cryptocurrencies or anything. When we were using the word crypto, it was always for a crypto algorithm or [creating], protecting data.

Jon McLachlan

Yeah, exactly.

David M'Raihi

Yes. And things obviously changed a lot in the last 10 or 15 years. But then, yeah, I mean, it got me kind of hooked, and actually started really working on that and completing my PhD with the same advisor, [Jacques Stern]. Some brilliant, brilliant [man]. And yeah, that's how I went into security, more by the, I would say, the research angle. And then things, you know, evolved. It's like everything, right? [From] research and moving to product and things like that.

Jon McLachlan

Yeah, you ended up in industry. You have an impressive set of companies on your LinkedIn.

David M'Raihi

So it was, and it happened in sort of a random way, literally. I was in Paris at the time, and I get a call from somebody I worked with in the past, and she was just like telling me, hey, I'm in San Francisco now, and we are starting, you know, this team, and I will need somebody to deal with security, and would you be interested? And I mean, at first I was like, well, I don't know exactly what you mean. But then thinking about it, it was more the kind of the opportunity of going to the US. And I mean, worst-case scenario, okay, you stay for a couple of years, you don't like it or it doesn't work out, then that's fine. I mean, you can go back. And it was in 1998. And now we're in 2022, and I'm still here. So actually, next year it will be 25 years in the Bay Area.

Jon McLachlan

Right, [unclear] almost [entering] a big 25.

David M'Raihi

Almost, almost, almost.

Jon McLachlan

That's a fun one. Amazing. I love it. Thank you for sharing. Was that, happened to be Augustin [Farrugia]?

David M'Raihi

No, no, but somebody who actually Augustin knew. But at the time, Augustin was still in Europe. I mean, he went to the US [unclear], but he was in Europe, and then I think he came back a little later. But yeah, that person had some connection with Augustin, which is funny, now that you mentioned that.

Jon McLachlan

I mean, I don't have too much of a connection to Augustin, except he was the one who found my resume out of a pile of stuff when I was applying, after graduate.

David M'Raihi

[He's] good at finding a good resume, so that's why he found yours, right?

Jon McLachlan

I mean, I think I'm [unclear], but I just count myself [to be] very lucky.

David M'Raihi

So you were just on the top of the pile, then he picked you up. But there we go.

Jon McLachlan

Yeah, I was just the last one to have submitted a resume, and it was a big stack of a thousand. And just, you know, at the last minute, I'm like, eh, whatever. Here we go. I'll submit this.

David M'Raihi

We'll never know. It's not that important.

Jon McLachlan

No, no, it's not. It's funny how the universe has a way of working. So, what would you say your best day has been throughout your entire career here in the Valley?

David M'Raihi

[unclear]. Yeah, I know, I know. And I was just trying to think about it. And I don't know if I can single out just one day, but I do think there were a couple of days, definitely. I mean, recently, when I think about it, I do think that what was happening at Symphony, when we went from just being a small kind of startup, [I'll say], this Perzo company, and then the combination and creating, and that day when we were sitting in that big room with all the bankers, and you have that kind of feeling that, okay, this is happening, you know. And that, I don't think I had that kind of [feeling] many, many times in my life, maybe like two or three times. That was definitely one.

Jon McLachlan

So tell the story a little bit. Like, okay, so there's a company named Perzo. And you guys were part of Perzo.

David M'Raihi

Yes, it was a company originally, I mean, it was really focusing on one aspect of communication. It was this idea of how can you make communication, secure communication, efficient and private between not just two parties, but multiple parties. And also, if you want to extend, you know, it's one thing to send a text message, but then what about you send a message and there is an attachment? Now you're talking also about what about the voice, what about the video? And I think when it started, the founders, they had many things in mind, but the kind of the pivoting was about financial application, because there is really a need for security and privacy and also compliance. And this is difficult. This is difficult. So not that many companies were interested in doing that. So I think moving from just having the technology to having a product and then having a company, that's what makes a difference. And by the way, that's what is really, really hard. And I think, personally, the reason why a lot of companies fail. And I've been [through] some failures as well. And you learn actually quite a lot, you know, when you fail. I think you learn even more when you fail than when you succeed.

Jon McLachlan

Right. But when you succeed, it just goes right to your head and you don't [learn] anything.

David M'Raihi

I think the issue is that sometimes, when, you know, when you succeed, and if you succeed very fast, I mean, all these things happen, and you don't really have the time to kind of digest them and understand what's happening, right? It's just, and look, I mean, sometimes you can be lucky, I mean, or just the stars are all aligned, and that can be also the fact that you have been working for 20 years on something, and then you're just at the right time with the right idea, and then things go very fast. So I think there are all sorts of configurations. But what is interesting is when you fail, you do have time. You always pause, because, pardon my French, but fuck it, did not work. Why did, what did I do, what did I do wrong? Or what my partners? Or, I mean, you start [unclear] or anything, but you're like, okay, let's try to understand. And usually there is always a lesson to be learned. And that's something that helps you, and you grow, and then the next challenge is [a bit] easier. And then, until you succeed, you succeed and [fail] again, then succeed. And I think it's kind of a mix of all these things. But you do learn a lot in these situations.

Jon McLachlan

If you bring that growth mindset. Yeah, some people don't.

David M'Raihi

That's right. Yeah, they don't. And also, I mean, going back very quickly on the product, technology, company. Many times there is a lot of confusion there, and people think that they have a company, and actually they just have a technology. Or they think they have a company, but actually it's a product, but there is no market for that product. All you need to do. And it is actually very, very hard to build a company that has a real purpose. And I think most of the times it's really because you have identified a set of problems, and you bring a solution that is a better solution than all the solutions existing, or just that there was no solution yet to that set of problems. And I think if you don't have that kind of angle, it's very hard. You can build something, but it's not going to be very solid.

Jon McLachlan

It's perhaps the difference between a company and a hobby.

David M'Raihi

Yeah, I think. And people don't really want to, you know, kind of [accept that]. I mean, they're like, hey, I'm building a company, and yeah, are you sure? And also, I mean, there is nothing wrong. You can have a great idea and build a technology that can be used by other people. And I think, to me, the open source movement and the capability of being able to offer some IP and some technology in that context, I think that's been very helpful, because then you can see that a lot of companies were creating, actually leveraging that technology. And it's, I mean, it's okay. Some people are very good business people. Some people are good researchers or scientific. You have also some talented engineers, but if you tell them, hey, okay, now go and market your stuff [or build a company], they will look at you like, not only I don't understand what you mean, but on top of that, I have zero interest in doing any of that. You know, it's boring for me.

Jon McLachlan

So it's sort of what makes teams stronger than the sum of their individual parts, when you bring people together and their strengths complement each other, right?

David M'Raihi

So, yeah. And it's valid for any team of any size, any organization. I think the diversity aspect, that's also something that I think sometimes is misunderstood. I mean, we use diversity these days in so many different ways. But I think it's really the diversity in terms of the way you approach things. And that's a result of your experiences and your background and who you are. So it maps pretty well with having people from very different backgrounds. And that does bring, I think, a lot of benefits to any team.

Jon McLachlan

So when you build teams, you look for a high diverse, highly diverse background as a candidate. So you're looking to [complement]. I mean, you're [gathering].

David M'Raihi

[No], because I'm more like, first, I try to find people. It's so difficult to find quality people. And because I don't know what's going to happen this year and next year. I think the economic situation and the world also changed with everything that has been happening in Europe and all the disruptions with COVID. Honestly, I mean, in the last two years, there have been so many impactful events that being able to predict and say, oh, next year we are going to have a big recession, or there is going to be even something worse than that. And now people, you know, they [can] decide. What I can observe, at least, I can see that a lot of people have simply decided, okay, wait a minute, am I happy? I mean, just, you know, very, some very basic questions that I think they did not have the time or they did not want to. And I see that even a lot more with young generations. And obviously it's, I mean, it's a little easier, right? You don't have a family, you don't, I mean, you might have some debts. So we will not get into that, because that's a very complicated topic. But I think that if you have a little more flexibility, of course you're going to say, hey, why would I do that job? Or why would I [be] there, or why would I [be] this way, or with this person? And I think there were a lot of questions around that. And people were accepting a lot more few years ago, and today they simply don't. Is it going to last? I think, of course, you know, it's like everything. Like, we went from an extreme where, you know, the power was on one side, and then it complete shifts, and it's like everybody can do pretty much whatever they want. I think things are probably going to rebalance a little bit. But I do believe that most people will always now have this kind of, you know, thought, at least in the back of their mind, is, hey, am I really happy? Do I want to do that?

Jon McLachlan

I mean, I think it's important that we lead lives that are deliberate and intentional and aligned with our own personal values. And sometimes, for some people, that means solving a very difficult open problem. Maybe for other people, it means, like, working to live, and working to provide for their families and the people that they love in their lives.

David M'Raihi

Oh, yeah. And there is, I mean, don't get me wrong, there is nothing wrong with that.

Jon McLachlan

But I really think that it's very reflective. Like, an interesting thing.

David M'Raihi

Yeah, exactly. It's just that I think a lot of people had kind of the time to think about it and realize, yeah, you know what, I do like being at home two days a week, because that way I can run some errands, and then I remove some stress from my life and my spouse, my family, or whatever it is. I can spend a bit more time with my family as well. But then I'm also happy to go into the office for three days, because I can see my team and my coworkers. And it's also a change of scenery. And I think for some people, it's just like, no, I don't want to do that at all. And other, I just wanted to be there all the time, because some people really enjoy being in an office. But until a couple of years ago, it was not even like an option, right? You would not go and say, you know what, I'm going to do this Monday and Tuesday.

Jon McLachlan

I think I'll be home every Monday and Tuesday, you know?

David M'Raihi

Like, that was unheard of.

Jon McLachlan

You're right. [And you'd], yeah, you'd be such an oddball if you had said something like that, you know.

David M'Raihi

Yeah, and also, you know, working remotely, literally in a different country and with a time difference. And don't get me wrong, I mean, there are many jobs, it doesn't work. It just is not going to work.

Jon McLachlan

It's true.

David M'Raihi

But there are also a lot of jobs, a lot of positions that are extremely flexible, and people can do that. And they also have probably the right mindset associated with that kind of job, and that works well for everybody. So, just think that it's, I think it's just more open. It's like the horizon is bigger. I don't know how to describe that, but yeah.

Jon McLachlan

So, does Rivian have a flexible work policy that facilitates this? You work from home most of the time as a CPSO, or?

David M'Raihi

So I try to go into the office on a weekly basis. I think when I think about that, I mean, at first I was like, okay, do I want to be there like four days or two days or one day, or is there a frequency that is best? And I think there is a value meeting with your teams. Do I absolutely need that? It's a good question. I mean, I think I can perform the vast majority of what I do remotely. But it's not just going into the office. For instance, I visited the plant a couple of times, and I do feel that going at least twice a year, or, again, I don't know what is the right rhythm, but you need to go there and to see and understand what it means making these vehicles. So, in our case, I think a company that is not building anything [unclear], that's probably slightly different, when you think of, I mean, [unclear] in security. Let's take, yeah, let's say that you're specializing in cloud security, right? And you are building all these things based on one of the [cloud platforms], whether it's, you know, AWS or GCP. I mean, at the end of the day, since you're building things in the cloud, what does it mean to, you know, be in the office? Except the connection with the other people. Right. And I noticed that [during] the kind of creative phase, right, being in the office and sitting together and going on the board and exchanging, it's easier.

Jon McLachlan

It's easier. Yeah, it's much faster and easier. You cut to the chase a little bit more quickly.

David M'Raihi

And you can see how people react.

Jon McLachlan

Oh, yeah.

David M'Raihi

Exactly. Exactly. And the other thing, it's the same thing when there is a crisis or a big, big problem. You can have a bunch of people in Zoom and that situation as well, and you walk through, but it's a little dry. And I think that being in the room can help. Now, when you're more in a kind of execution phase, and there is not really a lot of surprises, and people know what they have to do, the benefit of being in the office is disputable, because then you're interrupted, and some people also have a hard time to do context switching.

Jon McLachlan

I know, a person who's not, like, switching context. I publicly announced it to help reduce my context switching. I'm like, you know what? I am a very slow context switcher. I can do one thing very well, but I need the focus. Please make sure it's important if we're going to interrupt.

David M'Raihi

So, yeah, I do, I do, I do remember that. And, yeah, so in that case, being interrupted like 10 times during the day, I'm sorry, this is going to impact the productivity of the person in a very, very significant way. So. But look, I think it's like everything, right? I always thought that, you know, I love my wine, and [I like whiskey too], but I think it's like everything, you know, it's a little bit of everything in moderation, right? Including moderation. Including moderation. So sometimes you can go crazy, right?

Jon McLachlan

But sometimes, but also the crazy, yeah.

David M'Raihi

[unclear], yeah, no. And it's the same thing with, you know, everything, whether it's, you know, drinking or working very hard. It's just, I think you need to understand the Goldilocks zone.

Jon McLachlan

Just enough, you know, [that you're] happy and [you're] contributing and engaged, but not so crazy that you never want to go back again.

David M'Raihi

So, again, I think it's the kind of the how [you're] happy, as a kind of a tenet or something that you can really reference to. Because when you think about that, you know, there is this, I don't know if it's a definition, but, you know, there are happy [drunks], right? They are happy. And there are people, when they drink too much, they are terrible and they react. And that's right. But the reality is just like, look, if you realize that, you know, two glasses of wine make you good, but three makes you miserable, don't drink three. If you enjoy really working 10 hours a day in a specific context, well, try to do that. It doesn't mean that sometimes you won't have to do 12 hours, but there will be days where it will be only eight. So I think it's a question of really being able to decide. That's a big part.

Jon McLachlan

If you could pick out the worst day that you've had in your journey through Silicon Valley, could you share that with us? The worst. I know you're an optimist, and you can only see things in the positive light, but.

David M'Raihi

Well, honestly, if I look, I don't really necessarily want to speak [unclear] the details, no, but yeah. The worst days were always because of personal things, never because of work or company or project. Or if I really look at everything, no, it's like I focus on only work. But I think the reason I point that out is because, in the end, when you think very hard, personal stuff is what matters the most. It's some people, it can be friends, can be family, whatever it is, whatever you call it, you know, whatever you call it home, right? That's the thing that truly, truly matters, you know. In terms of really bad days, I would say that one day that was really hard, and I really did not know what was going to happen after that, is maybe three years after moved to the Valley, and 2001. And there is this day, you know, Enron goes bankrupt. And of course, I have shares in Enron. Like, I do not have a lot, but it's just enough. It was just, too, like, oh shit, you know, there is that, that is happening. But then you really see the market melting down. And the day that was very hard is, you know, the company I was associated with at the time, I mean, they decide to, like many other companies, they reduce the footprint, and doesn't make any sense to have a presence in the Bay Area. And it's just, and you see, you know, people, like, less and less people, and you really experience. And then you have this discussion with friends that tell, hey, you know, we are sitting in my office, and today there are only two people left. And I'm like, [who's] going to switch off the lights? It's him or you, or. And so, yeah, there was that really dark day of, you know, the kind of the fall of Enron, and then after that, really, the day when, okay, the office is closed, and what do I do? I mean, if I'm being, like, smart, I would just go back to France and, you know, the safety. But you didn't choose, and I decided to stay.

Jon McLachlan

You chose to stay. You wanted to [unclear]. Something, something about the [Bay] had sort of resonated with you.

David M'Raihi

Yeah, I think.

Jon McLachlan

The Silicon Valley bug.

David M'Raihi

I think it was a combination of, okay, that can't be it. You know, it's been three years. I mean, there must be [unclear]. The movie. Yeah. Yeah. I don't know, I don't know what it is.

Jon McLachlan

Did you feel empowered to sort of just move on and build your own story, or did you just want to see how this really ended?

David M'Raihi

I think, if I'm being honest, the first, like, couple of months, not having a job, not knowing exactly, not knowing anything at all. And also, I mean, you know, you do have some network, but it was very different at the time.

Jon McLachlan

I was about to say, you know, three years is not a very [important] [unclear] network.

David M'Raihi

Exactly. And also, it was like, you know, 2001, right? And [then] in the impact of social network and things like LinkedIn. I mean, you do not have all these tools. You barely have, you know, your phone and a few people email, and you would go and, you know, have drinks. But I was not, I was not scared or worried. It was more like, okay, what are the next steps, right? And it ended quite well, in the sense that I did know a few people, and then I connected with them, and there was this project starting. And so I could kind of rebound. And then, you know, after this one, there was another one. And things get, you know, much, much easier, because then you're connected and you have experience, and it's just, everything starts clicking. But yeah, that was [well], scary. A little scary, yeah.

Jon McLachlan

A little bit. Yeah, I could imagine, especially earlier on in your career, like, you have a little bit less of a safety net, and, yeah, a little bit smaller network, and you're in a different country, and.

David M'Raihi

And yeah, and then if it doesn't work out, then what do you do? I mean, do you really keep trying, or. But fortunately, I did not have to ask myself all these questions.

Jon McLachlan

So, yeah, no, that's great. I'm super happy for you that you decided to stay and that you found [unclear], you know, just a little bit. And here we are, and we got to meet and work together. It's from a place of deep gratitude and appreciation that I share with all of our listeners that David here has actually been my mentor through my career. [Well], you have helped significantly, and I have learned leadership styles and security technologies and how to interact with customers. And you gave me my first opportunity to really lead a team [unclear].

David M'Raihi

That's true. Yes, very true. I remember now. Yeah, that was how I [did it]. I remember [before].

Jon McLachlan

So, yeah, deep and authentic thank you.

David M'Raihi

No, you're way too kind. I learned a few things from you as well, so.

Jon McLachlan

I appreciate that. Well, I guess we both have gotten a lot out of each other, then.

David M'Raihi

I think so. I think so. [We still have to have] this vegan dinner when

Jon McLachlan

Well, that's probably [what] [happens] when [Baia] is done, you know, rebuilding after the fire or whatever [happened].

David M'Raihi

Did you set the kitchen on fire?

Jon McLachlan

I did not. I love that restaurant and absolutely did not [set] anything in that restaurant on fire. But [unclear] a very good restaurant. For those who don't know, it's a vegan, an all-vegan Italian comfort food, Michelin star quality restaurant.

David M'Raihi

Yeah, it's really good. It's an amazing restaurant.

Jon McLachlan

And even if you're not vegan, you'll have your socks blown off.

David M'Raihi

I confirm. I confirm.

Jon McLachlan

The non-vegan is confirming, and the vegan is most happy with that confirmation. So that's perfect. So if you look into the future, David, and I'll let you decide how far into the future you'd like to look, and you say to yourself, yes, that is a successful security industry, like, I am proud of what the security community has done. What is it that you're looking at, and how far into the future is it?

David M'Raihi

One trend that I see as a very positive trend is, and there are a bunch of companies in that space, but it's everything related to SecDevOps.

Jon McLachlan

SecDevOps.

David M'Raihi

Yeah, so security DevOps. And it just, I think, I mean, it's a bit of a broad term, but I like the space, because when you think about everything we do in security, there are always a lot of opportunities to, I would say, optimize certain things. And what we have been missing quite a lot is: how do we allocate our resources? How do we manage them the best? When do we need a tool? When do we need a human? And [particularly] in the space of vulnerability management, there are a bunch of players, obviously, and you can do some scans, you can do some pen testing, you can analyze the code [unclear].

Jon McLachlan

Yeah, it's kind of like what Lacework does, right? And then you get your [auditing] and [unclear] everything.

David M'Raihi

I think the space in terms of, in that case, it's, I mean, they do more than that. I mean, it's really the kind of the cloud security aspect and how you can automate [and] detection. But the very specific kind of angle where I am a company, I am starting to build some products, I don't have a security team. What can I do? What is, I would not say necessarily the bare minimum, but what would be a good solution for me, knowing that I can't hire 50 people? Right. But I still may not even be able to hire [one in] security.

Jon McLachlan

Yeah, you may not even be able to hire one yet.

David M'Raihi

Yeah, yeah. It's not necessary. So maybe we can have one kind of [unclear]. Yeah, I mean, security is a space, it's difficult to find people with experience and the right skill set and all these different things. And obviously that right price. But I really believe that at some point it was more like we are kind of brute-forcing everything. And, okay, I need 20 analysts, and I need a SOC, and I need this and that. And today I see a bunch of options where we can be a little smarter. And it doesn't mean that we don't need a team, but I see a lot of options where I can do today with 30, 35 people what 20 years ago would have required like a hundred.

Jon McLachlan

Wow. Is that because of all of the services, the SaaS services in the security space that have been bubbling?

David M'Raihi

It's a combination of things. I think people, again, in the industry, the skill set, the experience, the level of knowledge, and people are able to do multiple things. I think people were a bit more specialized. Today I see some engineers who can do infrastructure and operations and [code a little bit] and testing, and they have all these kind of [combination] of skills and knowledge in security. And when you think about that, I mean, security is kind of, I would not say it's a brand new industry, but when you think of, if you look just the encryption algorithms, right? I mean, the first time you had some public encryption algorithm, we are talking, you know, the 70s. And until the 70s, there was not a way, before RSA, and then some algorithms based on [discrete logarithms]. I mean, there are different variants, but basically the concept of having a public key that you use to encrypt, and then you have a private key that you use to decrypt. And then suddenly people start thinking, oh, that's nice, because I can also do signature, right? And so it started like a bit organic, and then you have a, but really, [unclear], a few people mid-70s. And of course, the military, government, that's a different story. But what was available, right, for the public.

Jon McLachlan

Yeah.

David M'Raihi

Yeah. And you see that in the 80s. And so you see a lot of security companies that also are going along with, oh, in the 90s, then people [start] having email. Then, since you have communication, then there are security issues associated with that. Yeah. Right. Yes, absolutely. And yeah. So I think it's also interesting to see that and be very humble, because a lot of people are saying, oh, but it doesn't work yet. And when you look at a car, and I [am in] a good place to see that, I mean, cars have been around for more than a hundred years, right? But the new kind of [new] [thing], when you think about an electric vehicle and the way they are built, and it's more like a connected object, and also all the technology around that, that's a shift again. And people need, you know, to really understand, hey, what do we need to do for this venture to be successful again?

Jon McLachlan

Do you think that we'll see the end or the break of RSA or ElGamal in our lifetimes?

David M'Raihi

I think there is a question of how this, [unclear]. Quantum cryptography is one thing, but all the algorithms that are like the post-quantum, and I do think that certain things are going to be more vulnerable. I mean, it's not clear to me yet, hey, does these things really work, and can you break things?

Jon McLachlan

So, just to be totally clear with all of our listeners, you're talking about quantum computing as being the thing that could potentially break RSA or ElGamal.

David M'Raihi

I think it's more like there was a lot of work in recent years to build algorithms that are based on problems that are very different. It's just that there was this big kind of question mark around, okay, the traditional problems, which are based on the number theory and things like factorization, because that's what RSA is about. Like, you have this [big] number, but there is no way to extract the two numbers that are kind of the product. Or you have a number, you can't extract [the log]. I mean, there are many, many, many things that were built around that.

Jon McLachlan

And then, because all of the internet security was built on those two problems, and the difficulty [of] two problems, right?

David M'Raihi

So, yeah. But the question was then, you know, okay, well, there are a lot of people that are thinking about what would happen in a world where these problems are not as hard as we think, you know, that they are. And I don't think that today, you know, we have to worry about it. Is this something that in 20 years? The reality is there are algorithms that have been designed, and there will be more in the future. And I think I'm more interested in the aspect that these new algorithms and new primitives give you in terms of functionality. And so, to me, when I look at, for instance, homomorphic encryption, and kind of this idea of, hey, now I have an encryption primitive, I produce some encrypted objects, and then I can operate on these encrypted objects without having to decrypt them. So then it's obviously super interesting, right? It's almost fascinating, because then you have a way of really maintaining the security of everything.

Jon McLachlan

So, yeah, that will be very exciting. That's what, like, Zama is working towards, the homomorphic encryption stuff. The Benoît from our Apple days, you remember?

David M'Raihi

Yes. Yes.

Jon McLachlan

So, let me ask you something completely different here, and a very open-ended question.

David M'Raihi

Sure.

Jon McLachlan

But when you're interviewing, you do a lot of team building there, Rivian and CPSO. I'm sure you get to meet a lot of very talented folks in the security world looking for their next adventure, professional adventure. When you're interviewing someone, what is it that you look for? What's your favorite interview question?

David M'Raihi

Look, the thing with interviews is most companies, I mean, there is a process, right? And because you want to be very fair and being able to compare the different candidates for specific positions. So obviously you're going to ask a set of questions, and you want also to be sure that you have people in the panel that are different. And so you have all these things. And it's complicated to be fair and balanced, but I think most companies do that reasonably well these days. But I always kind of [stray] away a little bit from the script. It's like, let's say I [have] 45 minutes. I will always use 30 minutes to be sure that, hey, I know that my talent team needs answers to these questions, and, you know, that's part of the evaluation. So I'm sure that I ask all these questions. But I always try to keep like 15 minutes out of 45, if it's an hour, like 20, I mean, I always [carve out some time] so I can ask more like personal questions, to try to figure out how the person, you know, think and click. And really, it's all about the fit with the company and the team, the organization values. Because when I look at all the people I worked with and people I hired, very rarely, when it was not working as expected, it was because of some lack in terms of technical skills or knowledge. You can always go, I mean, unless, you know, really make a big mistake, and you hire somebody to bake a cake, and is actually a sailor, and he doesn't know what a cake is. But most of these things, you think that the guy is, you know, is very good, let's say, in C language, and he's just good, but he wants to learn and improve. And that's something that you can easily fix over six months or a year.

Jon McLachlan

Sure.

David M'Raihi

But if there is a big discrepancy in the values and the way the person thinks, and, you know, it doesn't click, it doesn't click, they don't click with the team.

Jon McLachlan

It just might be a very good person, but just not the best fit, perhaps.

David M'Raihi

Exactly. Exactly. And that's when I'm being very, very, very careful. When I hesitate, it's very rarely because I'm not sure the person is technically good. And I mean, of course, that's part of the evaluation. But when I hesitate, it will, most of the times, like 80, 90% of the time, there is kind of this gut feeling: is that a good fit? And it goes both ways, you know, because there is, of course, you hire, yeah, because you hire the person, but also it's a big disservice if you hire somebody you think, and then the person is really trying their best, but it just doesn't work. And you don't want that to happen also, as well. So that's the reason I'm really very careful when it comes to that.

Jon McLachlan

That's good. No, I think cautious pays off, doesn't it? David, thank you so much for joining us on this episode of the Security Podcast in Silicon Valley. And thank you to all of our listeners who have tuned in for this episode. David, would you like to leave our listeners with any final words of wisdom? Advice that you wish you had for your younger self, perhaps?

David M'Raihi

Yeah, I think.

Jon McLachlan

For your older self.

David M'Raihi

As soon as you can, find your North Star, and don't listen to people if they tell you it's not the North. Just ignore them and go after it. The sooner the better.

Jon McLachlan

That's right. Don't hesitate. Be deliberate and intentional with [change].

David M'Raihi

And it's much better to make a big mistake than doing nothing.

Jon McLachlan

So, yes, you will never regret action.

David M'Raihi

[Nothing].

Jon McLachlan

You only regret not taking action.

David M'Raihi

Exactly. And it's exactly what happens, because, you know that I love to collect stuff, right? And it's always the books that you decide not to buy that you regret, never the books that you bought. So, but still, be careful. Don't buy any books.

Jon McLachlan

Be careful what you read, huh?

David M'Raihi

Exactly.

Jon McLachlan

David, [this has been] amazing.

David M'Raihi

Thank you, Jon. It was a pleasure. It was a pleasure. And looking forward to listening to the podcast again and again and again. And obviously seeing you soon.

Jon McLachlan

I look forward to it too. All right. Thank you, everyone.

David M'Raihi

Bye-bye, Jon.

Jon McLachlan

Bye-bye. Bye.

FAQ

Questions this episode answers

Who is David M'Raihi?
David M'Raihi was chief product security officer at Rivian when this episode aired, from January 2021 to December 2023. He holds a PhD in mathematics and computer science from the École normale supérieure, and he is first author of the HOTP and TOTP standards. He was chief security officer at Perzo and Symphony and head of product security at Pure Storage, and since April 2024 he has been global vice president and lead security at CARIAD, Inc.
What does a chief product security officer do?
A CPSO leads the security of the products a company builds. At Rivian, according to David's LinkedIn, that meant security in the factory, offensive testing of the vehicles, embedded security such as hardened control units and hardware security modules, and the infrastructure for key management and vehicle-to-cloud communication.
What are HOTP and TOTP?
Algorithms for one-time passwords. HOTP, RFC 4226 from December 2005, computes a code from a shared secret key and a counter. TOTP, RFC 6238 from May 2011, replaces the counter with the current time, by default a new six-digit code every 30 seconds. David is first author of both.
What turns a technology into a company?
A product with a market, built on a set of problems you have identified and a solution better than the ones that exist. Many companies fail at that step, David said, because they mistake a technology or a product for a company.
How can a company without a security team handle security?
David's answer was SecDevOps: decide when you need a tool and when you need a human. Scans, pen tests and code analysis cover a lot of ground, and he said he could do with 30 to 35 people what took about a hundred 20 years earlier.
Will quantum computers break RSA?
On the tape, David said certain things will become more vulnerable, and that nobody needed to worry yet. RSA rests on factoring, and Peter Shor's 1994 quantum algorithm factors integers, though beating classical computers may take millions of qubits. NIST released its first three post-quantum standards on August 13, 2024.
What is homomorphic encryption?
Encryption that lets you compute on encrypted data without decrypting it first. David called it almost fascinating. Jon mentioned Zama, and Benoît Chevallier-Mames of Zama was the guest on episode 20.
What does David look for in an interview?
Fit. In a 45-minute interview he spends 30 minutes on the questions his talent team needs answered and keeps 15 for personal questions about how the person thinks. Skills can be fixed in six months to a year, he said, and when he hesitates, 80 to 90 percent of the time it's about fit.
New episodes every other Tuesday

Got a story like this one?

We're always looking for practitioners who've been in the room when it mattered.