Designed and maintained CI/CD infrastructure across all ML services at Datature — accelerating release cycles and ensuring consistent, reproducible deployments to GCP. Managed Docker images, runtime upgrades, and shared library versioning across a multi-service ML platform.
Highlights
▸GitHub Actions CI/CD pipelines across all ML services — linting, testing, Docker builds, and deployment to GCP (Compute Engine, Cloud Run, GKE)
▸Multi-stage Docker builds to reduce image sizes; CUDA runtime upgrades (11.4 → 11.7) and Python version migrations (3.7 → 3.10)
▸Maintained PyPI package (Alchemy training library) and npm package (nexus-core-schema type library) with version consistency across services
▸KEDA authentication and production hardening for GKE-deployed video annotation services
▸Managed deployment configs across dev, staging, and production environments
Tech Stack
GitHub Actions Docker GCP GKE Cloud Run GCE KEDA PyPI npm CUDA