Data Science in Production
summary
Describes a working CI/CD pattern that stood up Google Cloud projects with terraform, used Vertex AI / Jupyter for experiments, versioned code in GitHub Actions workflows, pushed container images to Artifact Registry, and deployed models on Cloud Run. Advises creating a sandbox with wide-open permissions, log usage to narrow needed IAM roles, separate infra (infrequent changes) from apps (hourly to daily updates), and keep models, training, deployment and monitoring as containerized pipelines — approach proven at HCA to speed iteration.
AI-written summary · about
chapters (11)
- 0:00Introduction to talk
- 2:34Reality of data science in companies
- 7:09Tech stack overview
- 8:34Infrastructure as code basics
- 13:11Terraform and cloud setup
- 16:48GitHub workflow for infrastructure
- 18:33MLOps and development tools
- 21:02MLOps pipeline structure
- 24:11Application deployment flow
- 26:20Key takeaways and advice
- 27:45Q&A session
related talks
59mContinuous Intelligence: Keeping your AI Application in Production
35mHow to deploy machine learning models to production (frequently and safely)
19mData science for real problems
39mFrom Data Science to Production - deploy, scale, enjoy!
30mDevOps For Data Teams
22mDevOps in Data Science: What Works and What Doesn’t
59mDevOps for Data Science
42mCI/CD for Machine Learning