Hi folks,
We’re excited to share that we’ve open-sourced BharatMLStack — our in-house ML platform, built at Meesho to handle production-scale ML workloads across training, orchestration, and online inference.
We designed BharatMLStack to be modular, scalable, and easy to operate, especially for fast-moving ML teams. It’s battle-tested in a high-traffic environment serving hundreds of millions of users, with real-time requirements.
We are starting open source with our online-feature-store, many more incoming!!
Why open source?
As more companies adopt ML and AI, we believe the community needs more practical, production-ready infra stacks. We’re contributing ours in good faith, hoping it helps others accelerate their ML journey.
We’d love your feedback, questions, or ideas!
Top comments (0)