I'm a scientist turned research engineer with a background in machine learning and mathematical optimization, which I studied during my Ph.D. in Montreal.

I specialize in transforming cutting-edge research ideas into robust, high-performance, production-grade libraries and tools. My expertise lies in crafting reliable solutions that seamlessly integrate into diverse programming languages and environments.

Lately I have been involved in developing low level optimizations in Apache Arrow and significantly contributing to the xsimd generic SIMD library (check out out blog post on our improvements to integer unpacking). I also work in Jupyter, including Apache Parquet support, R support for CRDT and occasionally around the xeus stack (Jupyter kernel protocol in C++). I'm a main contributor to mamba, Conda-compatible package manager written in C++, and maintain many packages in Conda-Forge.

I thrive in rigorous development environments, where strong testing practices and static type-checking ensure code safety and maintainability. I believe in the power of building foundational, dependable tools that serve as universal building blocks for innovation.