Description
Industrialize the ML model lifecycle with MLflow, DVC, CI/CD, and monitoring.
Learning Objectives
- Describe the MLOps lifecycle and position each tool in the value chain.
- Configure an MLflow server for experiment tracking, model management, and registry.
- Version data and ML pipelines with DVC in conjunction with Git.
- Build a reproducible and parameterizable ML pipeline with DVC Pipelines.
- Expose a model in production via a REST API with FastAPI and Docker.
- Automate training, testing, and deployment via GitHub Actions (CI/CD/CT).
- Detect data drift and concept drift in production with Evidently AI.
- Monitor model performance in production with Prometheus and Grafana.
- Apply governance and reproducibility best practices on a real ML project.
Target Audience
Data scientists, ML engineers, data engineers, and Python developers aiming to industrialize their models and master MLOps practices in a professional environment.
Prerequisites
Proficiency in Python (functions, classes, virtual environments). Basic knowledge of Machine Learning (training, evaluation, scikit-learn). Basic use of Git and Docker.
Program Outline
Informations
Duration
3 jour(s)
21h
Tarif
Sur demande
Similar Trainings
IA & Data
Cloud Migration
2 jour(s)
Sur demande
IA & Data
AI Agents – Designing Autonomous Systems with LangChain and LangGraph – Advanced
3 jour(s)
Sur demande
IA & Data
European AI Act — Understanding your obligations and ensuring compliance — Beginner level
1 jour(s)
Sur demande
IA & Data
Data Analysis with Microsoft Power BI (4-167)
3 jour(s)
2200 € HT