Since January 2026, I’ve been building AI-native travel technology at Tribe & Trail. I raised capital and grew the business to $240k in monthly facilitated booking volume, alongside over $10k in monthly recurring revenue processed directly by us. The booking figure includes travel we facilitated without handling the payments ourselves.

Our flagship platform, Komo, is an agentic travel agent that has facilitated business travel for 16 Y Combinator startups. We’re expanding into consumer travel in Q4 2026; keep an eye out. Alongside Komo, I’ve run events including hackathons for Y Combinator companies, Google DeepMind, Firecrawl, Cactus Compute, Hugging Face and Nothing Phone, personally judging, securing sponsors and handling the full logistics.

At UBS, I provided quantitative sales teams with a weekly publication on institutional investor positioning in the markets, supported by data analytics and portfolio risk management solutions.

I built proprietary equity portfolio analytics tools comparing 230 institutional portfolios and presented the work to the Global Markets executive forum. In 2025, I ranked No. 1 in UBS Alpha Capture, Wall Street’s largest sell-side alpha capture platform, measured by absolute excess returns, returning $13m on a $25m portfolio.

My publications, sales commentary and quantitative work supported negotiations for multi-million-pound data subscriptions and access to our sales and trading commentary. I advised quantitative hedge funds on how the systems I built could inform trading within their pods and support their risk management.

As a junior analyst at UBS, I learned about AI and data intelligence while working on a content recommender platform. The platform shipped to more than 1,000 institutional clients and increased readership by 12%.

I also delivered bank-wide training for directors on the platform’s machine learning and automation enhancements, their wider benefits for clients and the regulatory environment.

My work also spans PRIZM, an agent orchestration layer; agent-based social simulation; routing between on-device and cloud models; and autonomous submersible navigation. Across these projects, I’m interested in how systems make useful decisions with incomplete information, and how we test whether those decisions are any good.

I work primarily with Python, SQL and Rust, alongside tools for forecasting, simulation and evaluating AI agents.