Welcome — What This Blog Is For

There's a lot that happens during research that never makes it into a paper.

A journal article usually gives you the final model, the final results, and a reasonably clean story of how you got there. What it rarely shows are the three approaches that failed before the one that worked, the strange dataset issue that cost you an entire week, or the moment when something finally made sense after staring at a reward curve for far too long.

Those experiences are messy, but they're often where the real learning happens.

That's what I want this blog to capture.

I work at the intersection of finance, machine learning, and reinforcement learning, with a particular interest in how these methods can be used for portfolio management. Through this blog, I'll write about the things I'm learning along the way — including the things that work, the things that don't, and the things that make me stop and question my assumptions.

What you'll find here

I'll mainly write about three things:

• Research notes — plain-language explanations of ideas from my research and ongoing work. I'll try to make them accessible to someone who understands finance but not necessarily machine learning, or machine learning but not finance.
• Market intuition — thoughts on what models and agents learn from financial markets, and how those learned behaviours compare with what we know about how markets actually work. Sometimes the two line up. Sometimes they really don't.
• Field notes — the practical side of doing research — working with financial data, building experiments, debugging models, dealing with imperfect datasets, and occasionally discovering that the backtest you trusted was lying to you.

If you're a researcher, a student exploring quantitative finance, or simply curious about what happens when you give a reinforcement learning agent a stock market and ask it to manage a portfolio, I hope you'll find something useful here.

I'm not trying to make the research sound simpler than it is. I'm trying to make it easier to understand — including the parts that usually get left out.

You can find my publications on the Research page, or jump straight into the first proper post: Why Deep Reinforcement Learning Might Change Portfolio Management.