ML Infrastructure Engineer @ Red Hat | PyTorch Contributor | India
I work on CI/CD infrastructure and developer tooling for machine learning frameworks. Currently building the Out-of-Tree CI integration for PyTorch's HUD — enabling third-party accelerator backends to surface their CI results alongside PyTorch's own test dashboard.
-
PyTorch OOT HUD Pipeline — ClickHouse queries, API endpoints, and React dashboard pages that display out-of-tree CI results on hud.pytorch.org. Co-authored the RFC and landed the full data pipeline (DynamoDB -> ClickHouse -> Next.js frontend).
-
PyTorch Core Contributions — Bug fixes and improvements across
torch.nn,torch.distributed, and test infrastructure. Notable contributions include fixing softplus numerical instability, adding convolution overflow checks, and fixing OOT backend registration.
Python C++ TypeScript React/Next.js DynamoDB AWS Lambda GitHub Actions Terraform PyTorch
Open to collaborating on PyTorch, ML infra, and open-source CI tooling.

