I’m a machine learning applied scientist at Secondmind, where I specialise on probabilistic models for time series analysis and forecasting. I received my PhD in Computing from Imperial College London. I actively do research around Gaussian processes (GPs) with particular focus on fast and efficient approximate inference. I am also interested in state-space models, stochastic variance estimation (heteroskedasticity) and optimising GPs with natural gradients.
A byproduct of my research is the creation of an internal library for linear time inference in GPs for time series, and, a model-based reinforcement learning investment strategy that runs on a fund.