Nicholas Riasanovsky
Staff Software Engineer at Meta
New York, New York, United States
Summary
🤩
Rockstar🎓
Top SchoolNicholas Riasanovsky is a Staff Software Engineer in New York with nine years of experience building compilers and high-performance parallel systems, currently driving Triton and GPU kernel performance at Meta. He has deep hands-on expertise in Python-native compilation and SQL optimization from senior engineering and technical lead roles at bodo.ai, and contributes to prominent open-source projects like Numba and Triton—adding robust datetime/timedelta NumPy ufunc support and improving AMD GPU backend scheduling. Nicholas blends production compiler engineering with diagnostic and testing rigor, having implemented unit-tested type handling fixes and scheduler diagnostics that reveal subtle recompile and ordering failures. His background teaching systems and computer architecture at UC Berkeley complements his engineering: he’s built a teaching compiler for RISC-V and mentored large student cohorts in OS and architecture courses. Colleagues describe him as a pragmatic problem-solver who surfaces hidden correctness and performance issues through precise low-level changes.
9 years of coding experience
6 years of employment as a software developer
Master of Science - MS Electrical Engineering and Computer Science, Master of Science - MS Electrical Engineering and Computer Science at University of California, BerkeleyEdison High School
Stackoverflow
71reputation
4kreached
5answers
0questions
Github Skills (21)
python10
numpy10
triton10
compiler-compiler10
compiler-optimization10
compiler10
numba10
amdgpu10
testing9
llvm9
mlr9
cprogramming-language8
parallel-computing8
c-language8
tkinter6
Programming languages (10)
Github contributions (5)
numba/numba
Jun 2021 - Jun 2022
NumPy aware dynamic Python compiler using LLVM
Role in this project:
Backend Developer
Backend DeveloperContributions:114 reviews, 106 commits, 23 PRs in 1 year
Contributions summary:Nicholas contributed to the Numba project by implementing and testing arithmetic operations between `datetime64` and `timedelta64` NumPy data types. They modified core files related to NumPy support and type handling, specifically focusing on ufunc implementations for datetime and timedelta operations. The user also addressed issues related to astype with literal strings, including modifications to array methods and type definitions. These changes involved adding unit tests and fixing bugs related to type conversions.
cudapythonparallelnumpynumba
triton-lang/triton
Nov 2024 - Apr 2025
Development repository for the Triton language and compiler
Role in this project:Back-end Developer
Back-end DeveloperContributions:12 reviews, 12 PRs, 35 comments in 4 months
Contributions summary:Nicholas primarily contributed to the AMD GPU backend of the Triton language and compiler. Their work involved debugging and optimizing the Ping Pong Scheduler, a technique used to improve performance, and improving the overall compilation process. They addressed issues related to reordering, re-compilation, and failures in the scheduler, adding diagnostic messages for debugging. The user also made optimizations and improvements related to the `is_within_2gb` function, impacting the interaction with buffer ops.
compilerprogramming-languagecode-generationtriton
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.