Balaji Veeramani
Software Engineer at Anyscale
San Francisco, California, United States
Summary
🤩
Rockstar🎓
Top SchoolBalaji Veeramani is a software engineer with eight years of experience based in San Francisco, currently building Ray at Anyscale. He specializes in distributed ML systems and backend engineering, contributing to Ray Train and RLlib while improving developer ergonomics and training callbacks. Balaji redesigned serialization/deserialization APIs for the SageMaker Python SDK and has a strong track record in model deployment and data handling. During internships at AWS he optimized distributed training on SageMaker and delivered a feature that reduced model hosting costs by 600% via Lambda deployment. Earlier research work includes maintaining NumS — a scalable NumPy implementation that outperformed Dask and Spark ML — and automating releases and CI/CD, reflecting a blend of rigorous academic foundations (CS & Statistics at UC Berkeley; Mathematics at UW–Madison) with production-grade open-source impact.
9 years of coding experience
2 years of employment as a software developer
Bachelor's degree, Computer Science and Statistics, 3.962 GPA, Bachelor's degree, Computer Science and Statistics, 3.962 GPA at University of California, BerkeleyMathematics, 4.0 GPA, Mathematics, 4.0 GPA at University of Wisconsin-Madison
Stackoverflow
36reputation
335reached
1answer
0questions
Github Skills (20)
json10
python10
machine-learning10
data-serialization10
numpy10
sagemaker10
serialization10
deserialization10
refactorings9
documentations9
documentation9
csv9
refactoring9
multiprocessing6
ray6
Programming languages (5)
Github contributions (5)
ray-project/ray
Dec 2021 - Jan 2023
Ray is an AI compute engine. Ray consists of a core distributed runtime and a set of AI Libraries for accelerating ML workloads.
Role in this project:
Back-end Developer
Back-end DeveloperContributions:1 release, 1461 reviews, 139 commits in 1 year 1 month
Contributions summary:Balaji contributed to the Ray Train and RLlib submodules, enhancing their functionality and addressing issues within the codebase. They clarified docstrings within the session class, converting TrainingResult to a dataclass, and monkeypatched environment variables in a callback test. Additionally, the user added a new PrintCallback for training results and made improvements to the documentation regarding code style.
pythonconsistsruntimetensorflowserving
aws/sagemaker-python-sdk
Jul 2020 - Aug 2021
A library for training and deploying machine learning models on Amazon SageMaker
Role in this project:ML Engineer
ML EngineerContributions:15 reviews, 36 commits, 66 PRs in 1 year 1 month
Contributions summary:Balaji's commits primarily focused on adding and modifying serializers and deserializers for handling data within the SageMaker Python SDK. They introduced base classes for serializers and deserializers and implemented specific implementations, including those for NumPy arrays, JSON, CSV, and JSON Lines formats. The changes involve updates to core components of the SDK related to data handling for model deployment and inference, demonstrating a strong understanding of data serialization and deserialization processes within the context of machine learning workflows.
pytorchsagemakerdeployingmxnetpython
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.