Production Deployment: Docker, APIs, and Scheduling
Moving a working notebook into a scheduled, monitored service
Expert
MLOps & Pipelines
1 lessons
Samuel Appiah Kubi
About this course
A pipeline that only exists as a notebook you run by hand isn't a monitoring system — it's a demo. This course covers what closes that gap: containerising a pygeovision pipeline properly, exposing it as a REST API, scheduling it to run automatically, and handling secrets the way a real deployment requires, all sized for resource-constrained infrastructure rather than assuming a data-centre budget.
What you'll learn
- Write a production-appropriate Dockerfile for a pygeovision pipeline
- Build a FastAPI endpoint for search, download and inference
- Schedule recurring monitoring with cron or Celery and Redis
- Deploy to a modest cloud VM rather than assuming unlimited infrastructure
- Handle credentials securely instead of hardcoding them
Requirements
- Advanced pygeovision (custom training recommended)
- Basic Linux command line
- Docker installed
Course content
A Production Dockerfile, Explained Line by Line
Preview
26 min
Free
- LevelExpert
- Lessons1
- CertificateYes
- AccessLifetime
Samuel Appiah Kubi
Instructor