Arjun Suresh
Senior Member Of Technical Staff at AMD
Milton Keynes, England, United Kingdom
Summary
🤩
Rockstar🎓
Top SchoolArjun Suresh is a Senior Member of Technical Staff with 11 years of experience specializing in code optimization, MLPerf benchmarking automation, and compilation workflows. Currently focused on MLCFlow and MLPerf inference at AMD and previously leading CM and MLCFlow development at MLCommons, he blends systems-level engineering with ML performance tooling. His open-source work includes integrating and improving the TVM backend and quantization support in MLCommons projects, driving reproducible, high-performance ML benchmarking across diverse hardware. A former researcher with a PhD background and roles at OctoML, cTuning, and academia, he brings both rigorous research discipline and hands-on production impact. Based in Milton Keynes, he also co-founded an education platform, demonstrating an appetite for building tools and communities beyond pure engineering.
11 years of coding experience
4 years of employment as a software developer
Master of Engineering (MEng), Computer Science and Engineering, Master of Engineering (MEng), Computer Science and Engineering at Indian Institute of Science (IISc)Higher Secondary, Higher Secondary at Christ Nagar Eng. Hr. Sec, SchoolB.Tech, CSE, B.Tech, CSE at College of Engineering Trivandrum
Stackoverflow
1reputation
0reached
0answers
0questions
Github Skills (15)
tensorrt10
machine-learning10
tensorflow10
onnx10
perf10
mlp10
python10
bert10
retinanet9
pytorch9
computer-vision9
tvm8
cuda8
onnxruntime8
image-processing7
Programming languages (11)
Github contributions (5)
mlcommons/ck
Nov 2021 - Jan 2023
Collective Knowledge (CK) and Common Metadata eXchange (CMX): community-driven projects to learn how to run AI, ML and other emerging workloads in a more efficient and cost-effective way across diverse models, datasets, software and hardware using MLPerf automations, CK playground and open challenges
Role in this project:
ML Engineer
ML EngineerContributions:203 reviews, 1510 commits, 842 PRs in 1 year 2 months
Contributions summary:Arjun's contributions focused on integrating and improving the TVM backend for the MLPerf inference application. They added support for various components, including the TVM framework and quantization, modifying the system's build process. The user also made changes to handle the inclusion of the dynamic batch size setting for performance and further improved and implemented various fixes.
pythonbenchmarkingmodularitycollaborationreuse
mlcommons/inference
Jan 2022 - Jan 2023
Reference implementations of MLPerf™ inference benchmarks
Role in this project:ML Engineer
ML EngineerContributions:148 reviews, 26 commits, 311 PRs in 1 year
Contributions summary:Arjun primarily contributed to the MLPerf inference benchmark implementations. They focused on modifying code to handle sub-millisecond target latencies in single-stream testing and updating the `run_local.sh` script. Furthermore, the user added and modified code related to the BERT model, including passing data/model paths via environment variables, adding options to skip accuracy verification, and overriding log paths. They also added support for multi-threaded preprocessing of images and implemented TVM backend.
implementationsbenchmarkingdeep-learningmlperfinference
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.