Isha Arkatkar
Software Engineer at Google
San Francisco Bay Area United States
Summary
🤩
Rockstar🎓
Top SchoolIsha Arkatkar is a software engineer with 15 years of experience building distributed systems and ML infrastructure, currently at Google in the San Francisco Bay Area. She blends rigorous backend and MLOps expertise—contributing to high-profile open-source projects like TensorFlow, Keras, and TensorFlow Federated—with a practical focus on distributed training, DTensor integration, and executor stability. Her background spans systems programming, parallel query processing, and production services from Teradata and Microsoft to startups, giving her a strong foundation across C/C++, C#, and Python ecosystems. A hands-on developer and lifelong maker, she approaches software with the same curiosity she brings to crafting and painting, often surfacing subtle correctness and documentation improvements that help large communities.
15 years of coding experience
4 years of employment as a software developer
Master of Science (M.S.), Computer Science, 3.56, Master of Science (M.S.), Computer Science, 3.56 at North Carolina State UniversityB.Tech., Computer Engineering, B.Tech., Computer Engineering at COEP Technological University
Github Skills (20)
c-language10
distributed-training10
python10
machine-learning10
distributed-systems10
keras10
mlops10
deep-learning10
tensorflow10
neural-network10
cprogramming-language10
ml9
distributed-computing9
deep-neural-networks9
automations8
Programming languages (3)
Github contributions (5)
tensorflow/tensorflow
Feb 2021 - Jan 2023
An Open Source Machine Learning Framework for Everyone
Role in this project:
Back-end Developer & MLOps Engineer
Back-end Developer & MLOps EngineerContributions:6 reviews, 144 commits, 4 PRs in 2 years
Contributions summary:Isha's contributions primarily involved updating and improving the documentation for the `tensorflow/tensorflow` repository, clarifying potential errors and edge cases related to specific operators. They also worked on lowering DTensorAllGather operations to CollectiveGatherV2 operations, adding an environment variable to disable this lowering. The user introduced tests for the EagerExecutor class, covering success and failure scenarios. These activities suggest a focus on both documentation and core framework maintenance, incorporating improvements related to distributed TensorFlow.
pythondata-sciencedeep-learningmlmachine-learning
keras-team/keras
Feb 2021 - Oct 2022
Deep Learning for humans
Role in this project:ML Engineer
ML EngineerContributions:7 commits in 1 year 7 months
Contributions summary:Isha made several contributions related to the TensorFlow/Keras library. These changes include modifying the code to handle parameter server strategies, integrating multi-process runners for distributed training, and updating the premade models test to include PS strategy combinations. Further modifications were implemented to allow the use of distributed datasets within tf.function and to update the optimizer to use replica_context.all_reduce. These changes suggest a focus on distributed training and optimization within the Keras framework.
pythondata-sciencedeep-learningneural-networksmachine-learning
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.