The paperwork, on request.
The short version is below and it covers everything except the contact details. For the full document, with grades module by module, referees and a way to reach me, send a request and I’ll reply personally. The only reason it is not a download is that it carries a phone number and an address. If you would rather just ask a question first, the same form reaches me and I’m happy to answer it.
Matthew Pearson
Computer Science & AI, University of Bath · Bath, UK · Open to 2027 graduate roles
mattpearson.dev · linkedin.com/in/matthewpearson274 · github.com/Matthew8747
MComp Computer Science and Artificial Intelligence
University of Bath. Integrated masters with placement year. Final year 2026/27.
- First class in all three years of study
- 82% dissertation
- A*A*A* at A-level in Maths, Computer Science and Physics, plus an A* EPQ
Where the work has happened.
A year on the Kontron Device Management product, across development and QA. Shipped Angular and TypeScript features from first bug fix to full sets, documented more than fifty defects, and built the Jenkins test runner that finally got dev-ops and the dev team running the same suite.
Built the visibility model behind NASCAR sponsorship value. A full Cup Series season of race results joined to Reddit, YouTube and news coverage, scored on published weights, then costed to find which sponsor buys the most exposure per dollar. Finished with a written report and a deck presented to the team.
Elected to the Students' Union Sports Executive to represent performance athletes. The seat carries their case into exec meetings, sits on the Blues Committee, works with the sports department on how performance sport is run, and drives student turnout at BUCS fixtures and showcase events.
Led campus tours and virtual open days for the computer science course, and mentored first-year students through their first months.
Microcontrollers and IoT security in Cambridge, where the interest in systems that have to be trustworthy started.
What it produced.
The case studies behind these carry the reasoning.
Set-piece simulator with a mandatory random-search baseline. On the study shipped with the repository, the optimised routine reached 6.66% mean xG per corner against a 1.04% baseline, with non-overlapping 95% intervals.
Production storefront and portfolio for a working painter. Stripe hosted checkout, Postgres, atomic stock so an original cannot sell twice, 101 unit tests and a security audit gating CI.
Website for the University of Bath Boat Club, built for a committee that changes every year. Editors publish a race report in under five minutes and it runs for about ten pounds a year.
Fifty-one participants across three sessions each. Designed the protocol, ran every session, and built the pipeline that reconciled six instrument streams onto one timeline. Graded 82%.
End-to-end analytics externship: four data sources across a 36-race season, a composite scoring model on published weights, and a ranking that held its order under four weightings and the full range of cost estimates. Delivered as a report and a presentation.
Led the user research and built the random forest recommender for an eight-person team project. A controlled study of 18 participants put our engagement ratio at 36% against the existing students' union site at 15%, with paired t-tests significant at p < .001.
What I reach for.
Competitive record.
- Henry Pearson-Gates award for outstanding A-level studies
- UKMT gold at every level
- British Physics Olympiad silver
- Medals at the Senior British Rowing Championships and at BUCS
- Rowed in the British Rowing Performance Development Academy