NicholaiĀ Tukanov
Software Developer at NVIDIA
Houston, Texas, United States
Summary
š¤
Seniorš
Top SchoolNicholai Tukanov is a software developer with eight years of experience focused on high-performance computing and low-level optimization, currently at NVIDIA after a Ph.D. candidacy in Electrical and Computer Engineering at Carnegie Mellon. He blends academic rigor with production-grade engineering, having designed and optimized matrix-multiplication microkernels and added POWER9/POWER10 support to the well-regarded BLIS linear algebra framework. A UT Austin computer science graduate, Nicholai has interned and worked at IBM and contributed measurable performance gains across low-precision formats (bfloat16, int16/8/4) in open-source HPC code. He enjoys making software go fast and is skilled at reworking code structure and test suites to improve maintainability and correctness. Based in Houston, he brings a pragmatic research-to-production approach to performance engineering on modern CPU/GPU platforms.
8 years of coding experience
2 years of employment as a software developer
Bachelor of Science - BS Computer Science, Bachelor of Science - BS Computer Science at The University of Texas at AustinMaster of Science - MS Electrical and Computer Engineering, Master of Science - MS Electrical and Computer Engineering at Carnegie Mellon University
Stackoverflow
1reputation
2kreached
0answers
4questions
Github Skills (18)
matrix-multiplication10
cluster-computing10
blas10
c1110
c1710
parallel-computing10
scientific-computing10
optimization10
bfd9
web-framework9
generics6
nasm6
arraylist6
compareto6
masm6
Programming languages (9)
Github contributions (5)
flame/blis
Feb 2019 - Jul 2021
BLAS-like Library Instantiation Software Framework
Role in this project:
Back-end Developer & Performance Engineer
Back-end Developer & Performance EngineerContributions:9 commits, 16 PRs, 9 comments in 2 years 4 months
Contributions summary:Nicholai contributed significantly to the BLIS library, primarily focusing on implementing and optimizing low-precision and high-performance matrix multiplication kernels. They developed and integrated POWER9 and POWER10 microkernels, including implementations for double-precision and various low-precision formats (bfloat16, int16, int8, int4). The user also enhanced the BLIS framework by adding support for POWER10, including test suites for assessing kernel performance and correctness. They reworked file structures to improve clarity and maintainability of the code.
software-frameworklinear-algebra-librarylinear-algebramatrix-multiplicationblis
nicholaiTukanov/blis
Feb 2019 - Dec 2019
BLAS-like Library Instantiation Software Framework
Contributions:789 pushes in 9 months
software-frameworkblasinstantiation
Find and Hire Top DevelopersWeāve 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.