Anshuman Suri
PHD Student at University of Virginia
San Francisco, California, United States
Summary
👤
Senior🎓
Top SchoolAnshuman Suri is a PhD student and applied machine learning researcher with 11 years of experience focused on adversarial ML, privacy, and security, currently based in Boston. He develops and evaluates defenses that aim to make models inherently robust, and has contributed practical attack and defense implementations to the widely used CleverHans adversarial library. His work spans academic and industry settings—from proposing novel algorithms like A-LINK and AMC (published at BTAS, T-BIOM, and IJCNN) to building ML features and AI-powered systems at Microsoft and research labs. Known for bridging theory and practice, he combines rigorous PhD-level study with hands-on engineering (fixing optimizer bugs, improving tutorials and docs) to make research reproducible and production-ready.
11 years of coding experience
3 years of employment as a software developer
Bachelor’s Degree (Hons), Computer Science, 9.45/10, Bachelor’s Degree (Hons), Computer Science, 9.45/10 at Indraprastha Institute of Information Technology, DelhiHigh School, Computer Science, High School, Computer Science at St.Columba's School, New DelhiDoctor of Philosophy - PhD, Computer Science, 4.0/4.0 (end of first year), Doctor of Philosophy - PhD, Computer Science, 4.0/4.0 (end of first year) at University of Virginia
Hindi, English
Stackoverflow
33reputation
11kreached
4answers
0questions
Github Skills (13)
machine-learning10
adversarial-attacks10
tensorflow10
python10
security10
benchmark9
benchmarking9
computer-vision6
deep-learning6
gpu6
datasets6
scikit-learn6
data-science6
Programming languages (9)
Github contributions (5)
cleverhans-lab/cleverhans
Feb 2018 - Jan 2021
An adversarial example library for constructing attacks, building defenses, and benchmarking both
Role in this project:
ML Engineer
ML EngineerContributions:10 commits, 18 PRs, 41 comments in 3 years
Contributions summary:Anshuman primarily contributed to the CleverHans library by implementing and refining machine learning attack methods. Their work involved fixing batch-size errors, adding support for custom optimizers, and incorporating checks for optimizer child classes. They also updated and maintained tutorial files, demonstrating a focus on applying and showcasing adversarial attack techniques within the context of MNIST and other machine learning tasks. The user also fixed documentation for attack parameters.
benchmarkingmachine-learningsecurity
iamgroot42/FormEstDistRisks
Sep 2020 - Aug 2022
Code for our work 'Formalizing and Estimating Distribution Inference Risks'
Contributions:39 commits, 8 PRs, 18 pushes in 1 year 11 months
inference
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.