Mechatronics BEng student at the University of Leeds (graduating 2028), currently on an industrial placement as an AI & Automation Engineer. I learn by building tools I actually use: quantitative finance research, embedded systems, and the glue between them.
| Project | What it is | Stack |
|---|---|---|
| systematic-trend-backtest | Pre-registered cross-asset trend-following backtest. Rules committed before any data was downloaded, sealed test set opened once, Deflated Sharpe and cost stress. Honest verdict: WATCH, not trade. | Python |
| ai-infra-research | Three stand-alone investment-analysis tools: a DCF/comps/Monte Carlo equity valuer on audited SEC EDGAR data, an AI-infrastructure supply-chain tightness map, and an institutional real-estate ROI model with an LP/GP waterfall. | Python, Streamlit |
| personal-engineering-environment | One C application bundling the mechanical, electrical and control-systems calculators I kept rewriting, plus data-analysis and productivity tools. Demo video. | C |
| Module | Project |
|---|---|
| ELEC2645 | STM32 games console: joystick, LCD, buzzer, modular C game logic |
| ELEC2301 | Machine learning on real datasets: modelling, evaluation and reporting in Jupyter |
| ELEC1301 | Wind-generation forecasting + Astrocrash: ARIMA time-series analysis and a refactored Pygame arcade game |
| Mechanics | Planar mechanism kinematics: scotch-yoke and slider-crank in MATLAB |
| Control | Maglev PID control: nonlinear modelling, state-space and PID tuning in Simulink |
- Research first, then build. Backtests are pre-registered, tests opened once, results reported even when they say "don't".
- Python (pandas, FastAPI, pytest), TypeScript (Next.js), C on STM32 and MBED, MATLAB/Simulink.
- Private repos cover a personal portfolio desk, financial models and a health-tech product in progress.