Alexander Yermolovich
Sr. Compiler Engineer at NVIDIA
California, United States
Summary
🤩
Rockstar🎓
Top SchoolAlexander Yermolovich is a senior compiler engineer with 7+ years building and hardening LLVM-based toolchains across top-tier silicon and software companies, currently driving middle-end compiler work at NVIDIA. He owns DWARF debug support in BOLT from his Meta tenure and has contributed production fixes and new DWARF features to the LLVM project, demonstrating deep expertise in debug-info, binary optimization, and compiler backends. His background spans GPU and integrated-graphics compilers (Intel, Samsung), ML compiler work on Glow at Cadence, and practical security and systems training from a UC Irvine masters. Known for bridging research-level compiler ideas with industrial constraints, he blends low-level implementation rigor with tooling and automation that scale across architectures and generations.
7 years of coding experience
20 years of employment as a software developer
BS, Computer Science, BS, Computer Science at University of California, BerkeleyUniversity of California, Irvine
English, Russian
Github Skills (14)
compiler-optimization10
compiler-design10
c-language10
bolt10
cprogramming-language10
dwarf10
llvm10
onnx9
debug9
cmake9
debugging9
quantization8
binarydiff8
python4
Programming languages (9)
Github contributions (5)
llvm/llvm-project
Jan 2023 - Jan 2023
The LLVM Project is a collection of modular and reusable compiler and toolchain technologies.
Role in this project:
Back-end Developer & Compiler Engineer
Back-end Developer & Compiler EngineerContributions:326 reviews, 1 commit, 115 PRs in 1 day
Contributions summary:Alexander's commits focus on enhancing the LLVM compiler infrastructure, specifically within the BOLT (Binary Optimization and Layout Transformation) tool. Their primary contribution involves modifying and refactoring helper functions within the DWARF (Debugging With Attributed Record Format) rewriter, addressing aspects related to debug information generation, particularly in .dwo files. They implemented updates for the DW_AT_comp_dir and DW_AT_dwo_name attributes, ensuring correct path information in generated DWARF debug sections. Additionally, the user addressed several issues, including fixing parent chains, handling DIE offset collisions, and adding support for new DWARF features.
compilerstechnologiesclangsubmittoolchain
pytorch/glow
Dec 2018 - Apr 2019
Compiler for Neural Network hardware accelerators
Role in this project:Backend Developer
Backend DeveloperContributions:8 commits, 62 PRs, 326 comments in 3 months
Contributions summary:Alexander made several contributions focused on improving the Glow compiler's functionality and stability. This included fixing a Windows-specific issue related to stack size, enhancing the framework with custom translation capabilities for Node IR, and addressing formatting issues. The user also added support for the LeakyRelu ONNX operator and disabled /fp:fast for VS builds, as well as included a feature for Quantization Schema option for RowWiseQuantization.
hardware-acceleratorscompilerneural-networkacceleratorshardware
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.