Gourav Roy
Senior Staff Engineer at Meta
Seattle, Washington, United States
Summary
👤
Senior🎓
Top SchoolGourav Roy is a Senior Staff Engineer based in Seattle with 10+ years building high-performance data infrastructure and ML systems for large-scale services at Meta and Amazon. He specializes in distributed storage layers, low-memory storage engines, pushdown computation, and production ML—work that drove a reported $100M training-data storage cost reduction and spawned a Data Tiering product built from scratch. Gourav blends deep algorithmic expertise (Robust Random Cut Forest, count-min sketch, HyperLogLog) with hands-on systems design, having implemented low-latency log-based engines, replication/snapshotting, and streaming aggregation features. He’s an experienced engineering leader who has led teams through complex deliveries and contributed to notable open-source ML tooling, improving RL training pipelines and memory-stable checkpointing in projects used by the community. An inventor and author with multiple patents and publications, he brings a rigorous, pragmatic approach to making large-scale, secure, and performant data systems.
10 years of coding experience
13 years of employment as a software developer
Bachelor's of engineering, Computer Science and Engineering, Bachelor's of engineering, Computer Science and Engineering at Birla Institute of Technology, Mesra
English, Hindi, Bengali
Github Skills (13)
amazon-sagemaker10
machine-learning10
deeplearning-ai10
mlops10
deep-learning10
tensorflow10
aws10
python10
reinforcement-learning10
devops9
jupyter-notebook9
dockers7
docker7
Programming languages (6)
Github contributions (5)
aws/amazon-sagemaker-examples
Nov 2018 - Mar 2019
Example 📓 Jupyter notebooks that demonstrate how to build, train, and deploy machine learning models using 🧠 Amazon SageMaker.
Role in this project:
ML Engineer & DevOps Engineer
ML Engineer & DevOps EngineerContributions:6 commits, 4 PRs in 3 months
Contributions summary:Gourav's commits primarily focus on integrating and configuring reinforcement learning (RL) training and simulation environments using Amazon SageMaker and AWS RoboMaker. Contributions include setting up RL training jobs with Coach toolkit and TensorFlow, incorporating RoboMaker simulation applications for both object tracking and DeepRacer scenarios, and configuring the necessary infrastructure. The user updated the codebase to utilize Coach 0.11.1 and made adjustments to memory usage and simulation parameters, suggesting a focus on optimizing the RL workflow.
pythonjupyter-notebooktrainingawssagemaker
IntelLabs/coach
Nov 2018 - Jan 2019
Reinforcement Learning Coach by Intel AI Lab enables easy experimentation with state of the art Reinforcement Learning algorithms
Role in this project:MLOps Engineer
MLOps EngineerContributions:14 commits, 1 PR, 7 comments in 1 month
Contributions summary:Gourav primarily focused on improving the stability and maintainability of the reinforcement learning training pipeline within the repository. Their contributions involved refactoring the checkpointing mechanism to avoid memory leaks in the rollout worker, which included modifying the graph manager and agent code. They also addressed issues related to data storage by enabling custom data store factories. Furthermore, the user introduced a fix to resolve memory leaks, ensuring the long-term operability of the system.
starcraftstate-of-the-artdeep-reinforcement-learningstarcraft2carla
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.