Advanced Machine Learning with Python

Syllabus

• Week 1

  AutoML and H20 Foundations

  • Navigating Your AutoML And MLOps Journey

    Discover why automated machine learning emerged and what it actually automates. You will compare the leading frameworks, set up H2O AutoML, and run your first automated model end to end.

  • Why Automated Machine Learning Emerged

    Understand the pressures that produced AutoML and the manual decisions it replaces. You will walk a conventional ML workflow, see where human judgement is repeatedly spent, and judge when automation is the right answers.

  • Evaluate AutoML Frameworks and Set Up H2O

    Survey the AutoML tooling landscape and the trade-offs that separate one framework from another.

  • Prepare and Validate Data for AutoML

    Learn systematic approaches to assess and clean data for AutoML. You will diagnose data quality issues before they impact model performance.

• Week 2

  AutoML Optimisation, Evaluation, and Productionisation

  • Engineer Features and Prevent Data Leakage

    Master feature engineering and learn to prevent data leakage, a critical mistake in ML projects. You will design feature sets that respect temporal and causal boundaries.

  • Configure and Control AutoML Model Search

    Guide and constrain H2O AutoML model search using search constraints and ensemble strategies. You will align model selection with your project's time, compute, and business requirements.

  • Evaluate and Compare AutoML Model Performance

    Learn when to use classification versus regression metrics and how to align metric choices with your business problem. You will discover why metric misalignment leads to production failures.

  • Optimise Hyperparameters and Refine AutoML Models

    Systematically explore the hyperparameter space to refine AutoML models and determine when to stop tuning. You will balance optimisation effort against diminishing returns using H2O tools.

  • Export and Prepare AutoML Models for Deployment

    Learn to export and serialise models for production deployment. You will understand deployment artefacts, versioning, and how to prepare models for real-time and batch scoring environments.

• Week 3

  MLOps Foundations, Versioning, and Pipelines

  • Track and Compare ML Experiments with MLflow

    Understand MLOps, experiment tracking with MLflow, and reproducibility practices. You will set up systematic experiment management for tracking parameters, metrics, and models.

  • Version Data and Reproduce ML Experiments with DVC

    Bring the same discipline to your data that version control brings to your code. You will use DVC to track dataset changes alongside commits, then reproduce an earlier training result exactly from its versioned inputs.

  • Automate ML Pipelines and Manage Model Versions

    Turn a working notebook into an automated workflow. You will build and run an Airflow pipeline, trigger it through GitHub Actions, and register the resulting models with the version and stage information.

• Week 4

  Production MLOps, Monitoring and Model Retraining

  • Serve and Deploy ML Models to Production

    Package a trained model as a service others can call. You will expose predictions through a FastAPI endpoint, containerise it with Docker, & deploy it to a cloud environment.

  • Monitor Model Performance and Automate Retraining

    Detect drift, monitor performance degradation, and automate retraining triggers in production. You will learn when to automate and when human decisions are needed for model updates.

Who will you learn with?