Ritika Borkar
Software Architect at NVIDIA
Beaverton, Oregon, United States
Summary
🤩
RockstarRitika Borkar is a software architect with six years of experience focused on performance engineering for deep learning systems, currently working as a DL Performance Architect at NVIDIA in Beaverton, Oregon. She combines hands-on ML benchmarking expertise with system-level architecture skills, having made notable open-source contributions to the mlcommons/inference repository that improved sample distribution, query generation, and performance-sample overrides for MLPerf inference benchmarking. Comfortable bridging research and production, she optimizes both tooling and test frameworks to reveal real-world model behavior under load. Her work reflects a practical knack for squeezing out performance while keeping benchmarks reproducible and auditable. Colleagues describe her as a pragmatic problem-solver who applies rigorous measurement to guide architecture decisions.
6 years of coding experience
Github Skills (8)
machine-learning10
benchmarking10
benchmark10
c-language10
cprogramming-language10
perf10
mlp10
python9
Programming languages (3)
Github contributions (5)
mlcommons/inference
Aug 2019 - Mar 2021
Reference implementations of MLPerf™ inference benchmarks
Role in this project:
ML Engineer
ML EngineerContributions:2 reviews, 26 commits, 12 PRs in 1 year 6 months
Contributions summary:Ritika primarily contributed to the `mlcommons/inference` repository by implementing and modifying features related to performance testing and benchmarking of machine learning models. Their work involved adding settings and modifying code to handle unique and same sample scenarios, and to override performance sample counts. They worked on code related to query generation, sample distribution, and the overall testing framework, enhancing the system's capability for evaluating model performance.
implementationsbenchmarkingdeep-learningmlperfinference
nv-rborkar/policies
Feb 2021 - May 2024
General policies including submission rules, coding standards, etc.
Contributions:64 pushes, 11 branches in 3 years 4 months
standardscontainer-coppoliciescoding-rulesrecall
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.