Hassan Kianinejad
Staff Software Engineer - ML Inference Performance at Twelve Labs
Houston, Texas, United States
Summary
🤩
Rockstar🎓
Top SchoolHassan Kianinejad is a Staff Software Engineer specializing in ML inference performance with a strong track record optimizing deep learning runtimes for GPUs and embedded systems. With roles at Twelve Labs, Intel (oneDNN), NVIDIA (cuDNN), and autonomous-driving startups, he blends low-level kernel tuning, OpenCL/CUDA expertise, and production-focused performance engineering. He contributed notable oneDNN GPU improvements—fixing matmul zero-point bugs and adding binary runtime scale support—demonstrating deep hands-on impact in a prominent open-source ML library. Trained in electrical engineering at University of Tehran and UCLA, his background in VLSI CAD and domain-specific computing complements his software work on hardware-accelerated inference. Based in Houston, he pairs a passion for outdoor team sports with a pragmatic, performance-first approach to shipping ML systems.
4 years of coding experience
7 years of employment as a software developer
BSc, Electrical Engineering, BSc, Electrical Engineering at University of TehranUniversity of California, Los Angeles
English, German, Persian
Stackoverflow
1reputation
0reached
0answers
0questions
Github Skills (16)
opencl10
oneapi10
performance-monitor10
c-language10
performance-analysis10
cprogramming-language10
gpu10
optimisation10
optimization10
deep-neural-networks9
neural-network9
deep-learning9
64-bit6
x86-646
x866
Programming languages (1)
Github contributions (2)
uxlfoundation/oneDNN
Dec 2021 - Jan 2023
oneAPI Deep Neural Network Library (oneDNN)
Role in this project:
Back-end Developer & Performance Engineer
Back-end Developer & Performance EngineerContributions:8 reviews, 60 commits, 2 PRs in 1 year 1 month
Contributions summary:Hassan primarily focused on optimizing the oneAPI Deep Neural Network Library (oneDNN) for OpenCL GPU architectures. Their contributions involved fixing bugs in the GPU OpenCL matmul implementation related to zero-point handling and adding support for binary runtime scale functionality. Furthermore, the user optimized the inference performance of the library by enhancing the ref implementation for inner product. The user demonstrated skills in optimizing and fixing bugs related to core kernels.
bfloat16sse41avx512openmpopencl
hkianine/oneDNN
Feb 2022 - Feb 2022
oneAPI Deep Neural Network Library (oneDNN)
Contributions:3 pushes in 7 days
deep-learningdeep-neural-networkoneapineural-networkonednn
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.