Dimitris Vardoulakis
Principal Compiler Engineer at NVIDIA
California, United States
Summary
🤩
Rockstar🎓
Top SchoolDimitris Vardoulakis is a Principal Compiler Engineer with 15 years of experience designing high-performance compilers, static analyses, and type systems for ML accelerators and large-scale software projects. Currently on NVIDIA’s XLA:GPU team, he brings deep HW/SW co-design experience from leading compiler support for multiple TPU generations at Google, including code generation for new ISAs and simulation-to-tapeout workflows. His work spans low-level performance engineering (notably contributions to TensorFlow and XLA) and high-level language tooling—he authored advanced type checkers and program analyses during his PhD and at Google’s Closure Compiler. A pragmatic tech lead and mentor, he combines research-grade program analysis (creator of CFA2) with production-facing optimizations like memory scheduling, tiling, and peephole passes to squeeze real hardware performance. He’s based in California and regularly contributes to prominent open-source ML and compiler projects, blending academic rigor with battle-tested engineering.
15 years of coding experience
11 years of employment as a software developer
Doctor of Philosophy (Ph.D.) Computer Science, Doctor of Philosophy (Ph.D.) Computer Science at Northeastern UniversityDiploma of Engineering Electrical and Computer Engineering, Diploma of Engineering Electrical and Computer Engineering at National Technical University of Athens
Greek, English
Stackoverflow
571reputation
17kreached
23answers
2questions
Github Skills (31)
debugging10
simplify10
debug10
javascript10
c-language10
type-check10
ui-components10
llvm10
gpu-programming10
google-closure-library10
tensorflow10
performance-optimization10
cuda10
xla10
compiler10
Programming languages (10)
Github contributions (5)
openxla/xla
Mar 2018 - Oct 2022
A machine learning compiler for GPUs, CPUs, and ML accelerators
Role in this project:
Back-end Developer & Performance Engineer
Back-end Developer & Performance EngineerContributions:21 reviews, 98 commits, 29 PRs in 4 years 7 months
Contributions summary:Dimitris's contributions center on optimizing the XLA compiler's performance and efficiency through code refactoring and algorithmic improvements. The user implemented and refined algebraic simplifications, specifically for broadcast operations and power functions, and improved the precision of the reduce operations in the interpreter. Furthermore, they added new opcodes for efficient broadcast implementations and fixed bugs related to memory allocation and aliasing, ensuring more accurate and faster compilation. The user's work improved the precision, speed and overall performance of the XLA compiler.
compilercommunity-drivenmachine-learningmodular
tensorflow/tensorflow
Mar 2018 - Oct 2022
An Open Source Machine Learning Framework for Everyone
Role in this project:Back-end Developer & Performance Engineer
Back-end Developer & Performance EngineerContributions:1 review, 107 commits, 1 comment in 4 years 7 months
Contributions summary:Dimitris primarily contributed to the performance and debugging aspects of the TensorFlow library, focusing on GPU-related optimizations and diagnostics. Their work included adding `ToString` methods to improve debugging, fixing typos, and increasing alignment for performance. They also addressed specific hardware configurations, such as Blackwell, and improved error handling in simulator setups.
pythondata-sciencedeep-learningmlmachine-learning
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.