Quantitative Analyst at UBS • PhD in ML-Driven Distributed Systems • Neuro-symbolic AI & Belief Revision Researcher
My academic foundation was built on graph theory and massively parallel computing during my Licentiate at the University of Buenos Aires. My PhD at the University of Westminster focused on response time prediction in distributed systems using machine learning — laying the groundwork for scalable, adaptive models.
Since I began studying belief revision, I have been deeply intrigued by how intelligent systems can maintain consistent knowledge under uncertainty and change. How can the AGM postulates be relaxed to preserve sufficient consistency while approaching human-like flexibility and robustness?
At UBS, I channel this expertise into UBS Quant Answers — a platform that empowers investors to interpret and manage complex market exposures with precision, transparency, and speed.
I guide, research, design, review, and develop UBS Quant Answers — a powerful and broad platform that helps investors interpret and manage their exposures and risks across a wide range of market factors and influences. UBS Quant Answers allows clients, from quant to fundamental investors, to tap directly into the innovation developed by the UBS Global Quant Research team.
The platform delivers portfolio risk forecasting, and a large suite of proprietary data and analytics modules. The Industry Network Intelligence constructs a handcrafted knowledge graph of analyst-attested relationships, serving as a living foundation for experimenting with graph-based models in finance. It provides the structured, expert-curated substrate that enables the integration of symbolic reasoning and neural learning—forming the starting point for a neuro-symbolic representation.
Data is delivered via API, Excel Add-in, SFTP, or UBS Neo, with rigorous data quality controls, automated checks, and 24/7 global support. This ongoing initiative drives innovation and efficiency in quantitative finance by combining advanced analytics, intuitive visualization, and scalable infrastructure — bridging theoretical computer science with real-world financial decision-making.
Join me at Imperial College London, South Kensington Campus (Imperial Business School Lecture Theatre LGS, Exhibition Road, London SW7 2AZ) for this free, in-person workshop exploring how AI is revolutionizing finance. My talk will delve into the evolving symbiosis between human cognition and machine intelligence, probing the trajectory toward technological singularity and its profound implications for financial systems, decision-making, and ethical AI deployment.
Secure your free spot now – spaces are limited!
Thesis: Response Time Prediction Models in Distributed Computing Environments using Machine Learning
Developed predictive models for performance optimization in large-scale distributed systems using supervised and reinforcement learning techniques.
Thesis: Graph Theory and Massively Parallel Computers
Explored graph algorithms on parallel architectures, focusing on scalability, load balancing, and communication efficiency in supercomputing environments.
Foundation in systems design, algorithms, software engineering, and formal methods.
Core Challenge: How to maintain a consistent, scalable knowledge base that adapts to new evidence without catastrophic inconsistency or prohibitive computational cost.
The AGM postulates offer theoretical perfection but fail at scale. My approach: graph-based, iterative belief propagation that finds “good enough” consistent states using local optimization — trading perfect rationality for practical, scalable reasoning.
This method naturally handles uncertainty, fuzzy logic, and probabilistic evidence. It bridges symbolic reasoning with neural learning by representing knowledge as weighted graphs where nodes optimize local consistency via expert-defined functions.
Key Advantages:
I'm always interested in discussing neuro-symbolic AI, belief revision, quantitative finance, or new challenges at the intersection of theory and practice.