AI

ESIA

School of Artificial Intelligence

Worldwide cohort
Students across time zones
HomeMaster’sTrainingsProjectsResearchBlogAboutContact
← Back to Trainings

MLOps Starter: FastAPI + Docker + CI/CD

Deploy an ML model as an API, containerize it, test it, and set up CI/CD basics for a production-grade delivery.

MLOpsFastAPIDockerCI/CDTesting

Duration

4 days

Format

Remote

Level

Intermediate

Key outcomes

  • Design ML API contracts (schemas, validation, errors)
  • Containerize correctly (reproducibility, image hygiene)
  • Add tests for preprocessing + inference + API endpoints
  • Create a simple CI pipeline with quality gates

Syllabus

Day 1 — API design for ML

  • Prediction endpoints and schema validation
  • Batch vs real-time inference patterns
  • Logging, error handling, status codes

Day 2 — Containerization

  • Dockerfile best practices (slim images, caching)
  • Pinning dependencies and reproducibility
  • Local build/run and environment parity

Day 3 — Testing

  • Unit tests for preprocessing and inference
  • Integration tests for API endpoints
  • Contract tests and example payloads

Day 4 — CI/CD essentials

  • Pipeline stages: lint/test/build
  • Artifacts and simple release strategy
  • Next steps: registry, environments, monitoring