My interest in machine learning sparked during my master’s thesis Computational Search for high redshift quasars which you can find here if you’re curious.

Since then my research focus has shifted to reinforcement learning, and with applications to trading and hedging. My main focus is on:

• risk averse reinforcement learning
• online learning (online convex optimization, experts, bandits)
• online planning (monte carlo tree search)

Reinforcement learning (RL) is one of the main areas of machine learning, the other two are supervised and unsupervised. What happens in RL is an artificial agent (a computer) is learning to achieve the objective you assign it by interacting with the environment, you can imagine it like trying to teach something to a dog, using rewards and penalties. This is the RL bible in case you’re wondering on where to start. If you’re a pro, and you’re curious on my work, checkout out my publications.

Instead the applications in which I am mainly focusing on are:

• quantitative trading (learning alpha generating strategies with low market correlation)
• portfolio optimization (finding the optimal multi-period portfolio allocations)
• market making (a dynamic policy which continuously prices an asset)
• optimal execution (dynamic optimal execution strategy)
• option hedging (can we beat the standard delta hedge?)

My Ph.D. dissertation can be found here.