Guillermo Lezama

Guillermo Lezama

PhD Economist | Data Scientist | Causal Inference & ML Systems

I am a data scientist and applied economist (PhD, University of Pittsburgh) specializing in experimentation, causal inference, NLP, and ML-driven systems. My work combines rigorous statistical methods with production-oriented tools — including large-scale text classification with LLMs and automated experimentation pipelines — to translate data into decision-ready insights.

Selected Work

NEW AI Voter Personas (ML + GenAI System) Clustered 51 ANES policy variables into 15 voter segments and built a structured LLM interface grounded in empirical distributions. View Project | Live Demo

Uruguayan Elections (Data Apps + Visualization) Built interactive applications to explore electoral data from 2004–2024, leveraging voter ID systems and precinct-level data. Featured in Semanario Busqueda. View Project | GitHub

Experimentation & Causal Inference (Amazon) Built RCT automation pipelines (ingestion, balance/SRM checks, treatment effects) and modeled and estimated causal lift from delivery-speed changes for investment modeling (using observational and experimental data).

More projects in Projects & Systems →

Working Papers

Valentina González-Rostani, José Incio, Guillermo Lezama (2025). Immigration Shocks and Politicians’ Rhetoric: Evidence from The Venezuelan Migration Crisis.

PDF

Dan Berkowitz, Guillermo Lezama, Claire Nackley (2025). Sanctions, Reputational Losses and Salience.

PDF

Guillermo Lezama (2024). Information About Corruption and Politicians' Proposals. Under Review.

PDF

K. Pun Winichakul (random order), Guillermo Lezama, Priyoma Mustafi, Marissa Lepper, Alistair Wilson, David Danz, Lise Vesterlund (2024). The Effect of Experimenter Demand on Inference.

PDF

Publications

Valentina González-Rostani, José Incio, Guillermo Lezama (2025). Social Media vs. Surveys: A New Scalable Approach to Understanding Legislators' Discourse. Legislative Studies Quarterly.

PDF Cite DOI

Guillermo Lezama, Henry Willebald (2020). Inequality in pre‐income survey times: a methodological proposal. Review of Income and Wealth, 66(4).

PDF Cite DOI

Contact