I walk to work in the morning, strolling past the historic buildings that surround our office. It’s cloudy today, but at least it’s not raining. I start thinking about the hedging method that I’ve been developing since starting at OxFORD. I joined recently, coming straight from a PhD in mathematics here in Oxford, but I was immediately given the freedom to choose a project that interested me.
Since I’ve joined, I’ve been working on a methodology to manage certain risk exposures, and have spent most of my time implementing, testing and improving it. I do this in Python, using both in-house and third-party libraries like cvxopt, and reading data from our API. It took me a few weeks to familiarise myself with the basics of our extensive infrastructure, but now I have a working algorithm and I’ve had some promising initial results. Today, I am comparing it against several of the approaches that my colleagues have developed over the past years.
Once in the office, I head to the refectory to grab myself a drink and some fruit. I chat about the weekend and some recent news here with some colleagues who have also decided to come in early today, and soon I make my way to my desk. I’m excited to see the results of the simulations that I left running on the research computers overnight. I’m still impressed with the accuracy and speed of the real-world simulator that has been built here over more than ten years – it is surprisingly difficult to do this properly. I analyse my results, and they’re better than my best so far! I quickly make two appointments with colleagues who have worked on similar projects, and one of them says he’s free so I go to present my approach to him right away. He makes several suggestions for improvements, and explains that his method is more applicable to a different type of strategy. However, he’s interested in my project and asks me to keep him updated on any further developments. It’s great to work with a set of people with such experience and domain knowledge.
I go out for lunch with friends from the office, and when I return, I focus on two side projects that I’ve been working on – one is evaluating an existing strategy and seeing if there are ways to enhance it. The other is constructing a portfolio from a new dataset of corporate events. The latter is greatly simplified by all the software tools that have been developed for analysing data and constructing strategies that trade any signals that the data might contain; my contribution is mostly the optimisation of this strategy’s parameters, and back-testing it rigorously. The actual implementation in C++ will be done by our software engineers, after which it will go live – my first actual contribution to the firm’s PnL!