AI Engineering (MLOps Foundations)

Learning Outcomes

✓ Design an MLOps pipeline (experimentation, CI/CD, deployment, monitoring)

✓ Containerize and deploy a simple model behind a REST API

✓ Operate basic Kubernetes-based inference and set up a CI/CD workflow

✓ Implement logging, metrics, and A/B tests; manage drift and retraining cycles

✓ Apply security and governance basics for data, models, and secrets

Curriculum

12-Week Curriculum

Week	Session A	Session B	Micro-lab
1	What is MLOps & why it matters	End-to-end pipeline overview	Sketch your pipeline for a toy use-case
2	Data pipelines & experiment tracking	Training pipeline patterns	Track runs with MLflow (local/Colab)
3	Deployment strategies (batch, online)	Building a REST API (FastAPI)	Serve a baseline model locally
4	Containers with Docker	Deploying with Kubernetes basics	Containerize API; run locally
5	CI/CD setup	Cloud services for MLOps (AWS focus)	GitHub Actions: test-build-push
6	Cloud MLOps (Azure & GCP)	Model monitoring & real-time tracing	Add basic logging & metrics
7	Feature stores & data management	Pipeline optimization & AutoML I	Register a feature set; try AutoML
8	Pipeline optimization & AutoML II	Cloud environment build I	Compare AutoML outputs & costs
9	Cloud environment build II	Monitoring & A/B testing	Canary/A-B plan for the API
10	Data & model security (secrets, SBOM)	Continuous learning & retraining loops	Secrets mgmt + drift detector sketch
11	DevOps in MLOps	Large-scale ops patterns & cost tuning	Add infra checks to CI
12	Large-scale ops patterns II	Final project workshop & wrap-up	Deploy a v1 (local or cloud)