My research is guided by the premise that economic and institutional systems are fundamentally dynamic, interconnected, and nonlinear. Firms, markets, and public institutions evolve through processes of persistence, adaptation, and structural rupture that cannot be adequately understood through static equilibrium frameworks alone. Understanding these processes requires models that capture time, network structure, and endogenous selection simultaneously.

To address this challenge, my research develops network-based econometric frameworks that integrate survival analysis, causal inference, and graph-based machine learning. A central contribution of this agenda is the Graph-Embedded Hazard Model (GEHM), a methodological framework that combines survival econometrics with graph neural network embeddings to model systemic risk and resilience in economic networks. By embedding economic agents within relational structures—such as financial exposures, supply chains, or institutional linkages—GEHM allows hazard models to capture not only individual characteristics but also the structural position of agents within complex systems.

This approach reflects a broader research objective: to bridge econometrics and modern machine learning in ways that preserve economic interpretability and causal reasoning. While many predictive AI models operate as black boxes, my work seeks to incorporate causal structures and network topology into statistical learning, enabling models that explain how shocks propagate, why institutions fail, and under what conditions policies become self-sustaining or fragile.

A second pillar of my research focuses on information structures and decision environments. Building on the theory of information comparison and information design, I study how informational changes interact with endogenous participation, selection, and institutional constraints. Recent work shows that classical welfare results—such as those derived from Blackwell comparisons of experiments—can break down once participation decisions depend on expected outcomes. These results identify structural conditions under which informational improvements remain welfare-robust, contributing to the theory of information design and economic decision-making under uncertainty.

Beyond methodological innovation, my research addresses a set of substantive questions about economic and institutional resilience:

• How do economic systems propagate shocks through networks of interdependence?

• What structural factors determine whether firms and institutions survive technological and political disruptions?

• Under what conditions do informational reforms or policy interventions generate robust welfare improvements?

• Can network-based machine learning models reveal the structural drivers of economic inclusion and development?

To answer these questions, my work combines econometric theory, network science, and computational methods, including graph neural networks (GNN), causal graphical models (GAN), and generative models for counterfactual analysis (GAT). These tools allow the simulation of alternative institutional or policy environments and provide a framework for studying systemic risk, technological diffusion, and structural change.

Ultimately, my goal is to develop a research program that advances a network-centric and temporally grounded economics, where institutions, firms, and policies are analyzed as evolving systems rather than static equilibria. By integrating econometrics, machine learning, and institutional analysis, this work aims to deepen our understanding of economic resilience and provide analytical tools that help policymakers navigate increasingly complex and uncertain environments.

Teaching is a natural extension of my professional and research journey. Over the past two decades, I have combined leadership roles in global firms, policy advising at multilateral institutions, and academic research in statistics and data science. This dual identity—professor and practitioner—defines my teaching philosophy: I aim to bridge methodological rigor with practical relevance, preparing students to use data and artificial intelligence not only as analytical tools, but as frameworks for responsible decision-making.

My teaching focuses on statistics, econometrics, and applied data science, with particular emphasis on how modern AI methods—such as survival models, graph neural networks, and machine learning architectures—can illuminate problems of risk, resilience, and innovation. I design courses around real-world dilemmas I have faced in practice, from modeling systemic risk and predicting firm survival to detecting fraud in financial systems. This case-driven pedagogy transforms the classroom into a laboratory where theory and practice constantly inform one another.

I have also taught extensively in executive education, working with managers, regulators, and policymakers across Latin America, Europe, and North America. Having lived the responsibilities of executives myself, I bring a unique ability to communicate complex technical concepts with clarity, translating them into actionable strategies. My global career has also shaped my teaching: by experiencing diverse institutional and cultural contexts, I equip students to see how methods operate differently across societies, markets, and governance structures.

Ultimately, my teaching philosophy is to train students as both analysts and decision-makers. By combining methodological depth, AI-driven approaches, and comparative insights, I seek to empower them to navigate the complex realities of business, policy, and research in a globalized world.