machine-learning-zoomcamp

Machine Learning Zoomcamp is a free four-month online course that teaches machine learning engineering from the ground up. It starts with core concepts and algorithms: regression, classification, decision trees, and deep learning, then moves through to putting models into production using Docker, FastAPI, Kubernetes, and AWS Lambda. All materials live in this GitHub repository and recorded lectures are available on YouTube.

The course is designed for people who already know how to program (at least a year of experience is recommended) but have no prior machine learning background. Deep learning sections that require more computing power use cloud resources rather than a personal laptop. The main tools used throughout are Python, NumPy, Pandas, scikit-learn, TensorFlow, and PyTorch.

There are two ways to participate. Self-paced learners can start at any time, work through the modules at their own speed, and build projects for a portfolio, but homework is not graded and no certificate is issued. Live cohort participants follow a fixed September-to-December schedule, submit homework for automatic scoring on a public leaderboard, complete two projects with peer review, and earn a certificate after meeting the requirements. The 2025 cohort starts September 15.

Each module has its own folder in the repository. Cohort-specific homework and deadlines are in a separate directory. Community support runs on the DataTalks.Club Slack workspace in a dedicated course channel, with announcements also posted to a Telegram channel.

Topics covered span the full engineering pipeline: exploratory data analysis, model building, feature engineering, evaluation metrics, model packaging, API development, container-based deployment, and scaling on cloud infrastructure.