
Project Topics
Ambition is cheap in October and expensive in March.

Every year a handful of projects die in March because the plan made in October needed a dataset that was never public, or a model nobody had time to train. We check the boring things first — availability, licence, hardware, hours — and then cut the idea into a core that must work and extras that are optional. Bring two or three ideas.
What you get
- Matched to skills you already have, plus one to learn
- Datasets and APIs confirmed available first
- Scope cut to a working core plus optional extras
- Something your supervisor will agree to supervise
- A title that describes what you will actually build
Related reading


Defending Your Demo
Ten minutes, one laptop, and a panel looking for the edges.
Read more →
Final Year Project
Proposal, build, report and defence — planned as one piece of work.
Read more →