Israel Ogbole
5 mins read
•
🎉 We are super excited to welcome Matti to zymtrace
Matthijs (Matti for short) is a software engineer based near Eindhoven, The Netherlands.
Before joining zymtrace, he worked in different areas of software engineering, with a strong focus on low-level systems and NT kernel development. He studied Computer Science at Radboud University in Nijmegen.
Much of Matti’s work can be seen in the open-source projects he contributes to, often focused on low-level system internals and performance. Examples include:
- EfiGuard (makes extensive use of zydis)
- The x64dbg debugger
- Anti-anti-debugging: ScyllaHide (user mode), TitanHide (kernel mode)
- ZeroInput: NT kernel driver developed for FireDaemon Zero
- Contributions to various Unreal Engine subsystems and general performance
Find him and his projects on LinkedIn or Github.
More about Matti
How did you get into software engineering?
This is a difficult question! I will say I’ve had an interest in computers from an early age, but it was quite a while later before I developed a serious interest in programming. I have a vague recollection of building the website for my primary school. This was long before today’s ubiquitous JavaScript frameworks, so this was all very bare bones.
Even though I did this and numerous other small programming jobs before I started my studies at university, at the time, I still had no real idea what I wanted to do! I was initially hesitant to study computer science at all, because of its notorious maths requirements, which was absolutely not my best subject in high school, to put it mildly. But looking back, I absolutely don’t regret my decision.
Odd as it may sound, I would say the most valuable course for me at university was functional programming. This was - and still is - taught at RU using a relatively obscure language named Clean that is mainly developed at the university itself. Clean is a pure FP language not unlike Haskell, but it deals with side effects such as I/O through uniqueness typing rather than monads. While I’ve never used a pure functional programming language after my studies, either professionally or for hobby projects, simply learning about the FP paradigm really changed the way I thought about writing software in any language.
Why zymtrace?
Working at zymtrace provides me with some opportunities that are a perfect match for what really drives me. I am particularly interested in (some might say obsessed with…) performance and efficiency, and as anyone who takes optimization seriously knows, you need to measure before you optimize. Because of this, I’ve long been a user and advocate of system profilers, and I couldn’t say no to the chance to work on one myself.
Second, working on zymtrace is a new challenge and learning experience for me in many respects. As a system profiler, zymtrace brings together many aspects of software and systems engineering that are not often found in combination in any single product: the tracer entry points, written in eBPF with all of its crazy restrictions, unwinding of both native frames and those of various interpreted languages, and of course the GPU profiler. On top of that, there is also the need to keep zymtrace itself performant while doing all of this!
Last but not least, the team at zymtrace consists of some of the smartest and most dedicated engineers I’ve had the pleasure of working with. This is important because it drives me to try and perform at the same level, which is a very high bar indeed.
If not in tech, what would you be doing?
I’d probably be a librarian! I love books (my parents are both librarians - this might explain). I also like to think I’m pretty good at writing, so that would be another option.
Teaching is another thing I really enjoy, though probably not as a job. But I do love answering interesting questions, whether they are programming related or not - especially if I don’t know the answer yet myself.
Which tools make your work easier?
Anything by JetBrains! In most programming languages, I wouldn’t be able to write a valid hello world program that actually compiles without having JetBrains tooling to do it for me.
Join us
We’re creating something special here, and we’re looking for people who care deeply about efficient AI workloads to build it with us. Join us.