Overview

Objective

Teach practical aspects of productionizing ML services — from collecting requirements to model deployment and monitoring.

[MANDATORY] Pre-requisites & setup

Note

TL; DR

• Docker Desktop
• Git
• Conda / Minconda / another Python environment manager
• Python 3.10
• Install ./requirements.txt

You can find all pre-requisites and setup instructions here.

Timeline

Course start: October 19th Course end: October 23th

Syllabus

Module 1: Introduction to MLOps

• What is MLOps
• Course overview
• Coding best practices
• Prerequisites and setup
• Running example: NY Taxi trips dataset

Module 2: Experiment Tracking

• Experiment tracking intro
• What is MLflow
• Experiment tracking with MLflow
• Saving and loading models with MLflow
• Model registry
• Practice

Module 3: Model Deployment

• Web service: model deployment with FastAPI
• Docker: containerizing a web service
• Practice

Module 4: Pipelines and Orchestration

• Tasks, Flows, Deployments
• From notebooks to Workflows
• Workflows orchestration with prefect
• Practice

[Module 5: Life Cycle Management]

• Model monitoring
• Model retraining
• Concept drift
• Data drift & data management

[Module 6: Project]

• End-to-end project with all the things above

Instructors

• DELATTRE Bruce
• BRITO Henrique
• BERTRAND Jules
• SERRA Luca