Boian Petkantchin
Senior ML Compiler Engineer at AMD
Santa Monica, California, United States
Summary
🤩
Rockstar🎓
Top SchoolBoian Petkantchin is a Senior ML Compiler Engineer with a decade of experience building high-performance, low-level systems for graphics, computer vision, and machine learning. Grounded in a mathematics degree, he specializes in lean C/C++ and assembly work that squeezes maximum performance out of constrained hardware and real-time pipelines. He has a strong track record across industry—from realtime video stabilization and digital-ink rendering to production ML tooling—and currently contributes to IREE-style MLIR-based compiler tooling and E2E training tests. Driven by curiosity, he pursues novel AI ideas (including genetic programming) and enjoys turning abstract math into practical, optimized implementations. Based in Santa Monica, he blends research instincts with hands-on engineering to ship robust, performance-critical systems.
10 years of coding experience
8 years of employment as a software developer
High school degree, Basics of computer science and programming., High school degree, Basics of computer science and programming. at 151 school “Sofia”Bachelor's degree, Mathematics, 3.04 GPA, Bachelor's degree, Mathematics, 3.04 GPA at Sofia University St. Kliment Ohridski
Bulgarian, English
Github Skills (10)
machine-learning10
jax10
mlr10
python9
tensorflow9
compiler-compiler9
compiler9
pytorch9
cuda8
runtimes8
Programming languages (13)
Github contributions (5)
iree-org/iree
Sep 2022 - Apr 2025
A retargetable MLIR-based machine learning compiler and runtime toolkit.
Role in this project:
ML Engineer
ML EngineerContributions:172 reviews, 103 PRs, 33 pushes in 2 years 7 months
Contributions summary:Boian primarily contributed to the development of end-to-end (E2E) training tests for machine learning models within the IREE framework. They implemented a test for an MNIST model trained using IREE Jax, verifying the accuracy of a training step. Their work involved generating test data, including the MLIR module, batch data, and expected optimizer states, using Jax and IREE Jax. Furthermore, the user also addressed and resolved an issue related to the model training process.
mlirspirvvulkantensorflowcompiler
sogartar/SHARK-Runtime
Nov 2022 - Mar 2025
Nod.ai 🦈 version of 👻 . You probably want to start at https://github.com/nod-ai/shark
Contributions:1 PR, 348 pushes, 107 branches in 2 years 4 months
pythonaspectsmachine-learningainod
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.