My research connects financial mathematics and machine learning with the design of practical, accountable investment systems.
I study stochastic optimal control problems arising in retirement income design, variable annuities, withdrawal guarantees, and path-dependent derivatives. This work includes analytical characterisation of optimal policies, weak convergence analysis, and regression-based Monte Carlo methods for high-dimensional control problems.
I develop portfolio construction and rebalancing methods that incorporate estimation uncertainty, market regimes, discrete trading constraints, and transaction costs. Current work ranges from robust optimisation to reinforcement-learning policies designed for production investment workflows.
This research uses structured and unstructured financial data to build interpretable signals for investment research. Areas include financial NLP, multimodal ESG assessment, macroeconomic state modelling, and validation frameworks that guard against overfitting and distribution shift.
I work on LLM-enabled systems for advisory, portfolio monitoring, and financial operations. The emphasis is on tool use, traceability, suitability and compliance controls, and the integration of AI agents with established optimisation and execution infrastructure.