Machine Learning Engineer course
from raw data to a served model.
A five-day, hands-on course in Cardiff or live online. If you live in Wales and have been made redundant or are under notice, the Welsh Government's ReAct+ grant can pay the full £1,500 fee.
- Duration
- 5 days
- Level
- Intermediate
- Format
- Cardiff or online
- Course fee
- £1,500
When your grant is approved before the course starts. ReAct+ can also help with travel and childcare costs while you train.
What the course covers.
Machine learning engineers turn data into models that make useful predictions, and then get those models into use. This course covers the full workflow: exploring data, building features, choosing and tuning models, and serving a model behind an API.
It is practical rather than theoretical. You work on realistic business data throughout, learn the mistakes that make models look better than they are, and leave with an end-to-end project.
- Pandas
- scikit-learn
- XGBoost
- PyTorch
- MLflow
- FastAPI
- Docker
Who it's for
- Analysts and Python users moving into machine learning
- Graduates in STEM subjects looking for an applied route in
- Developers who want to add ML to their skill set
Before you start
Basic Python and GCSE-level maths.
By Friday, you'll be able to
- Explore data and engineer useful features
- Build regression and classification models
- Evaluate models properly and avoid data leakage
- Train a simple neural network in PyTorch
- Track experiments with MLflow
- Package a model and serve it through an API
Five days, day by day.
Each day mixes short explanations with hands-on labs. Day five is for finishing and presenting your project.
- Day 1
The ML workflow
- Framing a business problem
- Exploratory data analysis
- Feature engineering
- Train, validation and test splits
- Day 2
Supervised learning
- Linear and logistic regression
- Decision trees and random forests
- Gradient boosting
- Choosing a model
- Day 3
Evaluation and tuning
- Metrics that match the problem
- Cross-validation
- Imbalanced data
- Hyperparameter tuning
- Day 4
Neural networks
- How neural networks learn
- Building a network in PyTorch
- Overfitting and regularisation
- When deep learning is worth it
- Day 5
From notebook to service
- Pipelines and reproducibility
- Experiment tracking with MLflow
- Serving a model with FastAPI and Docker
- Project demo
A churn prediction service
Predicts which customers are likely to leave, explains the main drivers and returns predictions from an API that another system could call.
Roles this course prepares you for
- Junior machine learning engineer
- Junior data scientist
- Analytics engineer
- ML operations assistant
What you leave with
A working project in your own GitHub portfolio and an Apex Training Wales certificate of completion.
How to get your place fully funded.
ReAct+ is a Welsh Government grant for people affected by redundancy. Your Working Wales adviser makes the application, and we give them the course details they need.
Check you're eligible
Aged 20 or over, living in Wales, with the right to work in the UK, and under notice of redundancy or made redundant in the last six months.
Speak to Working Wales
Call free on 0800 028 4844 or book an appointment online. Tell your adviser you'd like to take Machine Learning Engineer.
We send your course details
Contact us and we'll provide the course outline and fee your adviser needs for your ReAct+ application.
Start once approved
When your grant is approved, we confirm your start date. The £1,500 fee is covered, so there's nothing to pay.
Eligibility and grant amounts from Working Wales. Final funding decisions are made by the Welsh Government.
About the course and the funding.
How much maths do I need?
Will I cover deep learning?
Who is eligible for ReAct+ funding?
Will I have to pay anything?
Can ReAct+ help with travel or childcare?
What if I'm not eligible for ReAct+?
Ready to retrain in AI? Let's check your funding.
Tell us which course you're interested in and your situation. We'll explain the next steps and send your adviser everything they need.