An intensive six-week programme that takes you from ML fundamentals through production deployment, culminating in a real-world capstone project with industry mentorship.
The gap between building a machine learning model in a notebook and running one reliably in production is where most data science projects fail. Organisations invest heavily in model development only to find that their models never reach the users and systems they were designed to serve — or worse, they deploy without the monitoring and governance infrastructure needed to catch failures before they cause damage.
This six-week intensive bootcamp bridges that gap. Designed for experienced developers and data professionals who already understand programming and statistics, the programme focuses relentlessly on the engineering skills required to build, deploy, monitor, and maintain ML systems at production scale. Every week combines asynchronous online learning with hands-on in-person lab sessions where you work with real datasets, real infrastructure, and real deployment targets.
A model that lives only in a Jupyter notebook is a science experiment. A model that runs reliably in production, monitored and governed, is an engineering achievement. This bootcamp teaches you to build the latter.
The programme culminates in a capstone project where you design, build, deploy, and present a complete ML system addressing a genuine business problem — working under the guidance of an industry mentor who provides the kind of feedback and challenge you would receive in a senior engineering role. Graduates leave with a portfolio-ready project and the practical skills that hiring managers consistently identify as the hardest to find in ML engineering candidates.
This bootcamp is designed for technical professionals who have foundational programming and data skills and want to develop the specialised engineering competencies needed to build production ML systems. It is not an introduction to machine learning — participants are expected to arrive with working knowledge of core ML concepts.
The bootcamp opens by establishing a shared foundation in the machine learning concepts and workflows that underpin everything that follows. While participants are expected to have prior ML knowledge, this week ensures everyone operates from the same conceptual baseline and introduces the engineering perspective that distinguishes this programme from academic ML courses.
You will review the end-to-end ML workflow from problem framing through evaluation, with particular emphasis on the decisions that matter most in production: choosing appropriate evaluation metrics for business objectives, understanding data distributions and their impact on model behaviour, and recognising when a simpler model will outperform a complex one in deployment. The second half of the week focuses on feature engineering — the practice of transforming raw data into representations that maximise model performance. You will implement feature pipelines that handle numerical, categorical, temporal, and text data, and learn techniques for feature selection, dimensionality reduction, and feature store design.
Week two dives deep into supervised learning algorithms with an emphasis on scaling them to production datasets. You will implement and compare tree-based methods (Random Forests, Gradient Boosting, XGBoost, LightGBM), linear models with regularisation, and ensemble techniques — understanding not just how each algorithm works but when to choose one over another based on dataset characteristics, interpretability requirements, and computational constraints.
The engineering focus of this week centres on training at scale: distributed training across multiple machines, hyperparameter optimisation strategies (grid search, random search, Bayesian optimisation), experiment tracking and reproducibility, and cross-validation strategies that avoid data leakage. Lab sessions have you training models on datasets large enough to require careful attention to memory management, batch processing, and training time optimisation — the kind of practical challenges that academic courses rarely address.
This week covers the foundations and practical application of deep learning, moving from core neural network concepts through to architectures commonly deployed in production systems. You will build and train models using PyTorch, covering feedforward networks, convolutional neural networks for image and signal data, recurrent architectures and transformers for sequential and text data, and transfer learning techniques that leverage pre-trained models for domain-specific tasks.
The engineering emphasis is on the practical considerations that determine whether a deep learning model succeeds in production: GPU utilisation and mixed-precision training, managing training instability and debugging convergence issues, model compression techniques (pruning, quantisation, knowledge distillation) for deployment on resource-constrained environments, and strategies for handling the large datasets that deep learning models require. By the end of the week, you will have trained, optimised, and compressed a deep learning model that is ready for deployment.
Week four is where the bootcamp's engineering focus comes into sharpest relief. You will learn the principles and practices of MLOps — the discipline of deploying, monitoring, and maintaining ML systems in production — and implement a complete deployment pipeline from model serialisation to live serving.
Topics include model serialisation and versioning, containerisation with Docker, building REST and gRPC serving APIs, deploying to cloud platforms (AWS SageMaker, GCP Vertex AI, and Azure ML), CI/CD pipelines for ML (including automated testing of model performance), A/B testing and canary deployment strategies, and infrastructure-as-code for ML environments. Lab sessions guide you through deploying a model to a production-grade endpoint complete with authentication, rate limiting, logging, and automated rollback capabilities. You will also learn to design serving architectures that balance latency, throughput, and cost.
Deploying a model is only the beginning. Week five addresses what happens after deployment — the monitoring, governance, and maintenance practices that determine whether an ML system remains reliable over time. You will learn to detect and respond to data drift, concept drift, and model degradation using statistical tests and automated alerting systems.
The week covers building monitoring dashboards that track model performance, data quality, and serving health in real time; implementing automated retraining triggers and pipelines; model governance practices including model cards, audit trails, and approval workflows; and the organisational processes needed to manage a growing portfolio of production models. You will also explore the emerging regulatory requirements for model explainability and the tools (SHAP, LIME, Integrated Gradients) used to provide interpretable explanations of model predictions. By the end of the week, the model you deployed in Week 4 will have a complete monitoring and governance stack around it.
The final week is dedicated entirely to the capstone project. Working under the guidance of your assigned industry mentor, you will design, build, deploy, and present a complete ML system that addresses a genuine business problem. The capstone draws together every skill covered in the bootcamp — from feature engineering and model training through deployment, monitoring, and governance — and requires you to make the trade-off decisions that characterise real engineering work.
The week begins with a design review session where you present your system architecture and receive feedback from mentors and peers. You then have four intensive days of development, with daily stand-ups and mentor check-ins to keep you on track. The bootcamp concludes with a presentation day where each participant demonstrates their system to a panel of industry judges and fellow participants. Judges evaluate not only the technical quality of the system but also the clarity of the problem framing, the soundness of the engineering decisions, and the robustness of the monitoring and governance provisions.
The capstone project is the centrepiece of the bootcamp experience. It is designed to simulate the end-to-end workflow of an ML engineering project in a professional setting, complete with ambiguous requirements, imperfect data, and the need to balance technical excellence with practical constraints.
Each participant is matched with one of four industry mentors who provide guidance throughout the bootcamp and intensive support during the capstone phase. Mentors are practicing ML engineers and engineering managers at leading technology companies who bring current, real-world perspective to the programme.
Mentorship takes several forms throughout the bootcamp. During weeks one through five, mentors host weekly office hours where you can discuss curriculum topics, career questions, or technical challenges. During the capstone phase, your mentor provides one-on-one guidance on system design decisions, reviews your architecture, and challenges your engineering choices — replicating the senior engineer feedback loop that accelerates professional growth in industry.
Past mentors have come from organisations spanning fintech, healthcare technology, e-commerce platforms, and autonomous systems companies. Their involvement ensures that the bootcamp remains grounded in the practices and expectations of the current ML engineering job market.
Upon successful completion of this bootcamp, you will be able to:
This is an advanced, intensive programme. To keep pace with the material and contribute meaningfully to labs and discussions, participants must meet the following prerequisites:
A diagnostic assessment is provided upon registration. Candidates who score below the recommended threshold are offered a two-week pre-bootcamp preparatory module (included at no additional cost) covering Python for ML, linear algebra refresher, and Git workflows.
Throughout the bootcamp you will gain hands-on experience with the tools and platforms that define the modern ML engineering stack:
All cloud resources required for the bootcamp are provided through dedicated training accounts at no additional cost to participants. Setup guides and environment configuration scripts are distributed in the week before the programme begins.