Alex Zinenko
Paris, Ile-de-France
Summary
🤩
Rockstar🎓
Top SchoolAlex Zinenko is a compiler and ML systems engineer with 14 years of experience building high-performance compilation tooling and developer-facing interactive tools, currently serving as Principal Engineer at AMD after co-founding and heading research at Brium. He combines deep expertise in MLIR/LLVM, the polyhedral model and tensor compilation—contributing significant backend work to flagship projects like llvm-project, TensorComprehensions and IREE—to make software and hardware work more efficiently together. At Google DeepMind and Inria he led GPU compilation pipelines and production-oriented compiler research, bridging academic polyhedral techniques with real-world ML runtimes. An active LLVM Foundation maintainer and MLIR area chair, he influences upstream design (e.g., Transform dialect, GPU memory hierarchy and barrier removal) while mentoring cross-institutional research collaborations. Colleagues describe him as someone who obsessively simplifies complex IR and transformation logic, improving both generated performance and developer ergonomics.
14 years of coding experience
12 years of employment as a software developer
Doctor of Philosophy (Ph.D.) Computer Science, Doctor of Philosophy (Ph.D.) Computer Science at Université Paris Sud (Paris XI)Igor Sikorsky Kyiv Polytechnic Institute
English, French, Russian, Ukrainian
Stackoverflow
797reputation
77kreached
27answers
0questions
Github Skills (23)
c-language10
llvm10
machine-learning10
mlr10
code-generation10
compiler-design10
compiler-compiler10
gpu10
compiler-optimization10
cuda10
compiler10
cprogramming-language10
python9
parallel-computing9
computer-engineering8
Programming languages (10)
Github contributions (5)
facebookresearch/TensorComprehensions
Feb 2018 - Aug 2018
A domain specific language to express machine learning workloads.
Role in this project:
Back-end Developer
Back-end DeveloperContributions:378 commits, 109 PRs, 184 pushes in 6 months
Contributions summary:Alex primarily worked on improving the code generation and memory promotion aspects of the Tensor Comprehensions (TC) project, a domain-specific language for machine learning workloads. They refactored and optimized the code generation process, addressing issues in relation to CUDA, and ensured correct handling of strided accesses. The user also addressed several bugs and implemented tests related to memory promotion within the TC framework, improving the stability and performance of the compilation pipeline. Furthermore, they streamlined and refactored the codebase.
machine-learningdomain-specific-language
llvm/Polygeist
Sep 2020 - Oct 2022
C/C++ frontend for MLIR. Also features polyhedral optimizations, parallel optimizations, and more!
Role in this project:Back-end Developer
Back-end DeveloperContributions:7 reviews, 74 commits, 11 PRs in 2 years 2 months
Contributions summary:Alex focused on enhancing the `polygeist` project, a C/C++ frontend for MLIR, by addressing issues related to barrier removal and loop restructuring. Their contributions included implementing support for nested parallel constructs within the barrier removal process and resolving use-after-free errors within the loop restructuring passes. Additionally, they updated APIs to match recent LLVM changes and introduced a pass converting scf.for(scf.if) to scf.while for better optimization.
cppfrontend
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.