Machine Learning Researcher and Software Engineer who is a little obsessed with retrieval systems.
I train models to surface the right evidence, then build software that makes their behaviour easy to inspect.
📍 London, United Kingdom • 🔬 Neural Reranking, RAG, and Information Retrieval • 🛠️ Web, mobile, desktop, and browser tooling
- Machine Learning Researcher at Revax, sharpening the neural reranker behind a UK tax RAG workflow
- Hunting down hard tax queries and reranker failure cases, then turning what I learn into real model changes
- Finishing my MSc Applied Computational Science and Engineering at Imperial College London
My favourite question is not whether a system can produce an answer. It is whether the right evidence showed up before the answer was written.
| Project | What it is about |
|---|---|
| 🦅 Hawk Page Indexer | A published Chrome extension that indexes pages and ranks them with BM25F, built for an Amazon client team and recognised by Amazon Edinburgh engineers |
| 🎯 Adversarial Attacks on Ranking Systems | Research on how adversarial attacks bend neural and non neural retrieval pipelines |
| 🧬 Synthetic Epilepsy Labels | Generating more than 15,000 synthetic clinical cases from fewer than 100 real samples, with evaluation front and centre |
| 🧾 UK Tax RAG Neural Reranker | My current Revax work, where model quality meets product reliability |
Also fluent in RAG, neural reranking, recommender systems, parallel programming, CI workflows, and the occasional Haskell or Erlang detour.
- 🥇 First place at the Imperial College Fintech Hackathon with Project Wompoo, a product that shields seniors from phone call fraud, pitched over a weekend against more than 100 students from across the UK and Europe
- 🛡️ Built an autonomous tank agent at the Morgan Stanley GUTS Hackathon using a greedy strategy and a subsumption behaviour hierarchy
- 🎓 First Class Honours in Computing Science at Glasgow, with the Undergraduate Excellence Scholarship for all four years
- 🧑🏫 Class Representative, Peer Assisted Learning Leader, mentor, and demonstrator, because helping newer engineers find their footing is the fun part
Ask me about neural reranking, retrieval evaluation, or why a confident wrong answer is worse than no answer at all.
