Denys Shabalin
Senior Software Engineer at Google DeepMind
Zurich, Zurich, Switzerland
Summary
🤩
Rockstar🎓
Top SchoolDenys Shabalin is a Senior Software Engineer and PhD-trained compiler researcher based in Zurich, specializing in optimizing compilers and managed runtimes for high-level garbage-collected languages. He designed and implemented Scala Native—an LLVM-based AOT compiler and runtime—and co-founded Scalameta, the infrastructure behind widely used Scala tooling. At EPFL he built whole-program optimizations and a parallel Immix GC that delivered substantial performance and latency improvements, and at Google and DeepMind he has applied that expertise to automatic kernel scheduling, adversarial compiler test generation, and scalable formal verification for ML-generated code. His work blends systems-level performance engineering with practical developer tooling, and he has contributed performance and benchmarking support to prominent projects like tensorflow/swift-models. Known for turning deep research into production-grade systems, he combines academic rigor with a track record of leading open-source communities.
6 years of coding experience
10 years of employment as a software developer
Bachelor's degree, Applied Mathematics, Bachelor's degree, Applied Mathematics at National University of Kyiv-Mohyla AcademyDoctor of Philosophy - PhD, Computer Science, Doctor of Philosophy - PhD, Computer Science at EPFL
Ukrainian, Russian, English
Github Skills (10)
command-line-arguments10
swift10
benchmarking10
command-line-parser10
benchmark10
tensorflow10
performance-optimization10
argument-parsing10
json9
machine-learning8
Programming languages (6)
Github contributions (5)
tensorflow/swift-models
Nov 2019 - Jul 2020
Models and examples built with Swift for TensorFlow
Role in this project:
Back-end Developer & Performance Engineer
Back-end Developer & Performance EngineerContributions:12 commits, 24 PRs, 10 pushes in 8 months
Contributions summary:Denys implemented initial support for command-line argument parsing to enable running benchmarks with different settings. They introduced a modular structure, allowing for single command execution of both training and inference benchmarks and enhanced the output with descriptions. The user also focused on improving the performance and reporting of benchmarks by adding features like JSON output and incorporating additional metrics. Their work involved modifying the benchmark framework and integrating with the TensorFlow models.
swifttensorflowswift-for-tensorflow
google/swift-structural
Feb 2020 - Sep 2020
Contributions:3 reviews, 134 commits, 17 PRs in 7 months
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.