Projects
Side quests, academic deep dives and the occasional over-engineered weekend project. All open source.
Agents that argue about your holiday so you don’t have to.
A web app where multiple AI agents (flights, hotels, weather) talk to each other to plan a complete trip. A practical demo of multi-agent orchestration in travel.
View the code
Three architectures walk into a Twitter dataset…
Deep Average Networks, CNN+LSTM and Bi-Directional RNNs with attention, applied to author profiling on the PAN CLEF 2017 Twitter corpus. Built for my Deep Learning course at LSE.
View the code
What 80,000 listings say about London.
Data mining on Airbnb’s London listings: pricing trends, neighbourhood patterns and what actually drives a booking.
View the code
Statistically choosing where to fail your driving test.
A statistical comparison of pass rates between test centres to answer one question: where should a Londoner actually book their practical test?
View the code
Simulating an insurer’s worst year, 10,000 times.
Monte Carlo simulation of arrivals, departures and accident patterns for a car insurer, estimating downside capital risk with 95% confidence intervals.
View the code