Aart Bik
Distinguished Software Engineer at NVIDIA
California, United States
Summary
🤩
Rockstar🎓
Top SchoolAart Bik is a distinguished software engineer and compiler expert with over 25 years of experience building high-performance systems, from Intel’s automatic vectorizer to MLIR and sparse-tensor support at Google and NVIDIA. He blends deep systems-level knowledge—C/C++, assembly, SIMD, and parallel runtimes—with practical ML compiler work, contributing sparsity features to PyTorch, TensorFlow/XLA, and MLIR. His career spans optimizing compilers, large-scale graph processing, and VM/compiler work for Android and Dart, and he’s recognized with top industry awards for SIMDization and performance engineering. Comfortable across architectures (x86, ARM, GPUs) and languages (Fortran through Python/JAX), he is a strong proponent of fuzz testing and rigorous verification. Notably, he led MLIR sparsifier efforts at Google and now applies that expertise to accelerate sparse ML workloads at NVIDIA.
11 years of coding experience
28 years of employment as a software developer
PhD, Computer Science, PhD, Computer Science at Leiden UniversityMSc, Computer Science, MSc, Computer Science at Utrecht UniversityGymnasium Beta, Gymnasium Beta at Christelijk Lyceum, Gouda
English, Dutch
Github Skills (37)
pytorch10
c-language10
python10
llvm10
compiler-development10
sparse-matrix10
machine-learning10
vectorization10
mlr10
intermediate-code10
deep-learning10
tensorflow10
intermediate-language10
vms10
compiler-optimization10
Programming languages (10)
Github contributions (5)
llvm/llvm-project
Aug 2022 - Jan 2023
The LLVM Project is a collection of modular and reusable compiler and toolchain technologies.
Role in this project:
Back-end Developer
Back-end DeveloperContributions:414 reviews, 3 commits, 277 PRs in 5 months
Contributions summary:Aart primarily contributed to the MLIR (Multi-Level Intermediate Representation) compiler project, focusing on the `sparse_tensor` dialect. Their work involved modifying and extending operations related to sparse tensor manipulation, including `disassemble` and `assemble`. They also refactored and added test cases to validate the correctness of these operations, demonstrating expertise in compiler development and the intricacies of sparse tensor representations. Moreover, the user enhanced the existing infrastructure by adding new features such as ReLu operation during sparsification.
compilerllvmtoolchain
aosp-mirror/platform_dalvik
Jun 2015 - Aug 2017
Role in this project:Back-end Developer
Back-end DeveloperContributions:107 commits in 2 years 2 months
Contributions summary:Aart's contributions primarily involved fixing bugs related to data type mismatches and compiler warnings within the `libdex` library, specifically the `DexSwapVerify.cpp` file, which is responsible for verifying and swapping data in DEX files. They also reverted a previous change that restricted dexlist to 32-bit, indicating a focus on ensuring compatibility across 32-bit and 64-bit systems. The user also removed the dexlist code, which suggests involvement in library maintenance and dependency management.
Find and Hire Top DevelopersWe’ve analyzed the programming source code of over 60 million software developers on GitHub and scored them by 50,000 skills. Sign-up on Prog,AI to search for software developers.