Lecturer in Credit Risk and Fin Tech
Roles and Responsibilities
Co-Organiser of the annual QFRA: Quantitative Finance and Risk Analysis international symposium (Zakynthos 2026, Corfu 2025, Santorini 2024, Crete 2023, Kos 2019, Mykonos 2018)
Research Director supervising PhD students in Artificial Intelligence, East Asian Economies, Statistics, and Econometrics; MSc students in Quantitative Finance, Risk Management, and Credit Scoring
Course Designer & Instructor:
• Time Series Forecasting (Postgraduate)
• Data Driven Business Insights (Postgraduate)
• Statistical Learning in Banking (Postgraduate)
Background
Stavros specialises in building quantitative systems that move from academic research to production deployment in financial markets. His commercial work focuses on AI-powered investment infrastructure for wealth managers and family offices — autonomous systems that collect market intelligence, generate systematic signals, and support portfolio decisions.
His applied projects have delivered significant revenue uplift for FinTech and banking clients, including credit risk optimisation at scale (44M+ users), AI-driven supply chain optimisation that materially reduced stockouts and increased profitability, and the development of proprietary algorithmic trading platforms backed by substantial private R&D funding.
On the research side, Stavros has created two original methodologies. AION is a geometric time series forecasting framework that outperforms 130+ benchmark models — including 710M-parameter foundation models — with zero pretraining and sub-second CPU inference (paper in preparation). Pattern Causality is a methodology for detecting hidden causal dependencies in financial markets, published in two papers in the Proceedings of the National Academy of Sciences (PNAS) with National Academy member Prof. H. Eugene Stanley. Both are available as open-source Python and R packages.
He holds a PhD in Applied Mathematics from the University of Liverpool (Best PhD Thesis 2020, Visiting Scholar at Caltech).
Research Interests
AION — Geometric Time Series Forecasting
A forecasting framework grounded in Takens' embedding theorem that recovers hidden attractor geometry to forecast complex systems. Benchmarked against 130+ baselines across 8 scientific domains including financial time series. Outperforms 710M-parameter foundation models with zero pretraining and sub-second CPU inference. Paper in preparation.
Pattern Causality — Hidden Causal Dependencies in Financial Markets
A methodology for detecting nonlinear causal interactions — including "dark causality" — in financial markets and complex systems. Published in two PNAS papers with NAS member Prof. H. Eugene Stanley. Available as open-source Python module (C++ performance core) and R library (CRAN). 110+ GitHub stars.
Plouton — AI-Driven Multi-Asset Trading System
A production-grade systematic trading platform covering equities, crypto, and derivatives. Features a medallion-architecture data pipeline, ML-driven signal engine with multi-timeframe feature engineering, and autonomous execution capabilities. Designed for institutional-grade signal generation across asset classes.