Course Objective & Scope
Across enterprises, many accurate models remain trapped inside Jupyter notebooks: they cannot be retrained cleanly when data changes, break when dependencies shift and take weeks to turn into a service. This hands-on programme walks step by step through turning a single notebook prototype into a reproducible, testable and deployable software artefact.
Modularising code (uv/Poetry, Ruff, pytest), defining data and pipeline stages with DVC, logging runs and models in MLflow, wrapping inference in FastAPI and Pydantic, containerising with Docker and automating checks with GitHub Actions are combined into one cohesive project. Participants leave with a working MLOps reference repository they can adopt as an internal team template.
Who Should Attend?
- Data scientists and ML engineers who want to ship their own models to production
- Software developers helping data science teams bring code up to production standards
- Technical leads establishing a standard project template for growing AI teams
- Prerequisite: Experience building ML models in Python and basic Git proficiency
Key Takeaways
- Refactor exploratory notebook code into a configuration-driven, testable Python package
- Define data preprocessing, training and evaluation as a reproducible pipeline
- Track experiments in MLflow and wire the promoted model from the Model Registry into a service
- Package a production-ready prediction API with FastAPI, Pydantic schema validation and Docker
- Automate code linting, data/model unit tests and container builds in GitHub Actions
Curriculum
From Notebook to Production Code: Project Layout and Dependencies
- Standard MLOps repository layout, deterministic environments with uv/Poetry and Hydra/YAML configs
- Refactoring feature engineering and training logic into modular, reusable functions
Reproducible Pipelines and Experiment Tracking
- Defining datasets and multi-stage training pipelines with DVC (dvc.yaml)
- Logging metrics, hyperparameters and model artefacts to MLflow Tracking and Model Registry
Testing and Quality Gates for ML Code
- Unit tests (pytest), data schema validation (Pandera/Great Expectations) and training smoke tests
- Model regression tests checking latency, memory footprint and baseline accuracy thresholds
Serving Models with FastAPI and Docker
- Lifespan model loading, Pydantic input/output schemas and health/metrics endpoints
- Building compact, secure container images with multi-stage Dockerfiles
CI/CD Automation and Delivery
- GitHub Actions workflows for linting, unit testing and pipeline verification on pull requests
- Pushing version-tagged images to a container registry and emitting structured inference logs
Other Courses in This Category
-
MLOps Fundamentals
MLOps training covering machine learning model lifecycle management, containerization with Docker and Kubernetes, CI/CD automation with MLflow/Kubeflow, and model drift monitoring.
- Live Course
- 6 modules
-
Systematic experiment design, hyperparameter optimisation with Optuna and Ray Tune, metric/artefact standards, reproducibility and comparative setup of tracking tools (MLflow, W&B, Neptune), for data science teams.
- Live Course
- 5 modules
-
MLOps – Advanced Level
Enterprise-grade MLOps covering model drift monitoring, A/B testing, canary deployments, automated retraining, CI/CD pipelines, Kafka/Spark Streaming, and Kubernetes orchestration.
- Live Course
- 6 modules
-
MLflow Tracking Server architecture, PostgreSQL/S3 setup, autologging, custom model packaging with pyfunc, MLflow Evaluate (classical ML and LLMs) and Model Registry integration, for teams standardising on open-source MLflow.
- Live Course
- 5 modules
-
W&B Experiments, Sweeps, Artifacts (data and model lineage), visual error analysis with Tables, LLM tracing with Weave and collaborative Reports, for deep learning and generative AI teams.
- Live Course
- 5 modules
-
Synchronised versioning of large datasets, labels and model weights alongside Git using DVC, LakeFS and Iceberg/Delta time-travel, with end-to-end lineage for teams building auditable ML systems.
- Live Course
- 5 modules
Get in touch
