Even Quantum Computers Can’t Crack This Encryption

That’s why in 2016 NIST launched a competition looking for new encryption approaches resistant to quantum computers. In 2022, the agency announced the first round of winners, which included the CRYSTALS-Kyber scheme coauthored by Bos, who is technical lead of the post-quantum cryptography team at NXP Semiconductors in Leuven, Belgium.

Today, Bos is focused on integrating the algorithm into NXP’s portfolio of embedded hardware products, which includes chips for credit cards, contactless payment terminals, Internet of Things devices, and cars.

As someone who loves solving puzzles, Bos was well suited to a career in cryptography, he says. The fact that he can help make the world a safer place while doing something he enjoys is a massive bonus.

“If doing this research was practically useless, I would probably still do it,” he says. “But it’s super cool that you can work on interesting math puzzles, and then, in the end, it will have a very positive impact on everybody around you.”

Discovering Cryptography

Bos grew up in a small town close to Haarlem in the Netherlands and was fortunate to have an early introduction to technology. His father worked at a bank and had a desktop computer at home. Bos started using it to play video games but became fascinated by the underlying technology and quickly picked up coding skills. By the age of 15, he was already doing freelance programming jobs for various companies.

In high school, he learned about more formal computer science topics, such as algorithms and computational complexity. He found these subjects fascinating and in 2001 enrolled at the University of Amsterdam to pursue a bachelor’s degree in computer science. After graduating in 2004, he stayed on to get a master’s degree in grid computing, which he completed in 2006.

“If you design a cool algorithm, it could end up in the crypto library of Microsoft.”

While working on his master’s, Bos says he found himself drifting toward algorithm design and more math-heavy computer science, but he was also eager to continue working on practical problems. Then he discovered cryptography, which bridges his interests. “It’s really at the intersection of engineering, computer science, and mathematics,” he says.

This realization prompted Bos to apply for a Ph.D. program in the lab of renowned cryptographer Arjen Lenstra at the École Polytechnique Fédérale de Lausanne in Switzerland. Bos was accepted and started in 2007, just as the lab began investigating the use of unconventional hardware—such as gaming consoles—to do cryptanalysis, the process of breaking encryption.

His Ph.D. project involved building a cluster of more than 200 PlayStation 3 consoles and using it to crack a popular encryption scheme based on the mathematics of elliptic curves. The consoles’ multicore processors used the Cell architecture developed by IBM, Sony, and Toshiba, which was well suited to running lots of computing processes in parallel, as is required in cryptanalysis.

Learning About Lattices

During his Ph.D. studies, Bos worked on a summer project with another acclaimed researcher, Peter Montgomery, who was at Microsoft Research at the time. The pair clicked, says Bos, and he was invited to become a postdoctoral researcher in Montgomery’s lab in Redmond, Wash., after completing his Ph.D. in 2012.

Shifting from academia to corporate R&D was an invaluable experience, says Bos, as he got to see how research is translated into real-world products. “That was really motivating,” he says. “If you design a cool algorithm, it could end up in the crypto library of Microsoft, which then gets used by hundreds of millions of people worldwide.”

While at Microsoft, Bos began working on an emerging approach known as lattice-based cryptography, which depends on the mathematics of vectors in a grid. These schemes were promising because they could be used for both quantum-secure encryption and fully homomorphic encryption, a technique that makes it possible to carry out computations on encrypted data without first decoding it.

But after two years in the United States, Bos and his wife wanted to be closer to home. So in 2014, he took a job as a cryptography researcher at NXP and moved to Belgium. He joined the company’s innovation team, which comes up with features for products several years down the company’s product road map.

By then, advances in quantum computing made it clear that more secure encryption approaches would be important, says Bos. So working in collaboration with researchers from Arm, IBM, SRI International, and various universities, he helped design the lattice-based CRYSTALS-Kyber encryption scheme, which was submitted to NIST in 2017.

From Cryptography Research to Products

Since then, Bos has focused on implementing the algorithm in NXP’s embedded hardware. Lattice-based encryption requires considerably more memory than older approaches do, which makes it tricky to run on smaller chips like those found in ID cards or IoT sensors. His team had to make changes to the underlying mathematics of the algorithm and redesign it to run on these specialized chips.

Bos says his job has evolved significantly over the years. He’s gone from conducting pure research to leading a team and collaborating closely with other departments to translate their innovations into actual products. He had to work hard to develop the skills to act as an interface between engineers and business-focused teams, he says.

Being team leader is a high-pressure role, he says, because NXP has to stay ahead of the curve when it comes to post-quantum encryption. The chips it designs are used at the start of a long supply chain, he explains, as they have to be integrated into larger systems made by component manufacturers. Those systems are then sold to device makers or automotive companies that have to integrate them into the final products.

Each of those steps can take years, says Bos. That means NXP’s chips need to be quantum secure now so that the end users can meet government-recommended deadlines to migrate to post-quantum encryption by the early-2030s.

A Friendly Field

One of the things Bos likes the most about cryptography is that the field is relatively small and welcoming. “Everybody’s super friendly,” he says. “If you go to a crypto conference, the big names, the folks who really invented crypto in the ‘70s, they still come to these events and you can meet them in person.”

The size of the field also means cryptography experts are in short supply, Bos adds, so it’s a discipline with great career prospects. While many roles require strong math skills, there are plenty of opportunities for those with a more conventional computer science background, and companies are always looking for electrical engineers to build cryptographic hardware.

A postgraduate degree in one of those fields is a bonus, but companies like NXP do a lot of internal training so it’s not strictly necessary, Bos says. Taking cryptography or security courses online or while at university can be a great differentiator, he adds. But most important is the right attitude. “You just need to be motivated and curious and willing to learn,” says Bos. “I think these are really the biggest factors.”