An end-to-end AI fashion assistant that transforms personal wardrobes into digital assets. It features automated background removal, multimodal AI analysis for clothing categorization, and RAG-based outfit recommendations using vector search.

A quantitative study on the UK housing market using Explainable AI (SHAP) to analyze the relationship between EPC ratings and property values. It uncovers an inverted correlation in London where historical character outweighs modern energy efficiency.

A production-grade FastAPI backend for a live iOS dating app, featuring multimodal AI pipelines for social interaction analysis, chat orchestration, and automated conversational intelligence using Google Gemini.

A privacy-first, on-device AI agent implementing a Hybrid Neuro-Symbolic architecture. It bridges the gap between LLM natural language generation and deterministic logic to provide hallucination-free assistance using Apple's MLX framework.

An advanced probabilistic forecasting system for the UK National Grid, leveraging Google's Temporal Fusion Transformer (TFT) to predict electricity demand with high interpretability and quantile-based uncertainty quantification.

A large-scale quantitative analysis of the S&P 500 volatility surface, featuring a custom-engineered vectorized Implied Volatility solver and an empirical study of risk metrics (Vega) during the COVID-19 market crash.