Learning path
AI Engineer
Applied machine learning: build, evaluate and serve models in production systems.
- Level
- Intermediate
- Estimated time
- 8-10 months
- Required builds
- 6 projects
- Guided study
- ≈ 280 hours
What you will be able to do
- 01Implement and debug training loops rather than copy them.
- 02Design evaluations that catch leakage, drift and slice-level failure.
- 03Serve a model behind an API with latency and cost budgets.
- 04Communicate model limitations honestly to non-specialists.
Curriculum
4 modules, in order
Stages build on one another. Nothing is optional, and nothing is repeated for length.
- M170h
Maths & Python for ML
Vectors, gradients and probability, implemented in NumPy before any framework appears.
- M290h
Modelling
From linear models to transformers: architecture choices, training dynamics and debugging.
- M360h
Evaluation
Splits, leakage, calibration, per-slice error analysis and honest baselines.
- M460h
Deployment
Batching, quantisation, caching, monitoring and the cost model of inference.
Core skills
PythonNumPyPyTorchEvaluationServing
Prerequisites
- Python fundamentals
- Linear algebra basics
- Comfort with the command line
How you are assessed
- Reproducible notebook submissions with fixed seeds
- Three model builds benchmarked against a shared baseline
- An evaluation report reviewed by peers
Roles this leads to
AI EngineerML EngineerApplied ScientistStart the AI Engineer path
Talk to us about entry level, timing and how the reviews work.