Neil Rajgor

Quant teams face a constant challenge: pricing and risk models grow ever more complex, with high-dimensional PDEs, stochastic control, and nested simulations demanding both accuracy and speed. Recent advances—like Prediction-Enhanced Monte Carlo, deep learning PDE solvers (PINNs), and sophisticated interpolation techniques—are enabling solutions to problems once deemed intractable.

The biggest IT story in the last decade is how falling compute costs have enabled new modeling techniques that were previously impractical, such as machine learning and deep neural networks. But what is the state of ML and AI (even transformers?) for time series, and where are things heading?