Pattern Recognition: Understanding market patterns via alternative macroeconomic context
Leveraging work on visual representations of time-series for modelling cross-asset interactions.
ML for Inverse Problems: Calibrating a heston model with IV surfaces
Trainining a very simplistic model to generate synthetic volatility surfaces and using that to train a surrogate model to invert the heston model.
Embedding Predictive Models for Neutron Reflectometry Curve Analysis
Worked on applying Embedding Predictive Models + Recurrent Neural Network for the time-evolution of NR Curves @ Oakridge National Labs.
Pattern Recognition: Visual trends analysis of FX prices
Studying spatial-temporal / geometric representations of FX dynamics.