Summary
Felipe Polo is a statistician and machine learning researcher with nine years of experience applying data analysis, econometrics, and causal inference to public policy and education. Currently a PhD student in Statistics at the University of Michigan and a Student Researcher at Google DeepMind, he focuses on computational statistics, probabilistic ML, and Bayesian inference to make AI evaluation more efficient. He has interned at the MIT-IBM Watson AI Lab, where he developed statistical models to extract LLM skills and studied scaling laws and plateauing benchmarks for model evaluation. Felipe blends applied impact work—policy evaluations and education research—with deep technical rigor from his mathematical statistics training at IME-USP and international experience at the University of Tokyo. He also founded a data science study group to mentor peers and translate statistical ideas into practical projects, highlighting his mix of leadership, teaching, and research. Notably, he brings an economist’s causal perspective to ML evaluation, enabling insights that go beyond standard performance metrics.
9 years of coding experience
4 years of employment as a software developer
Mestrado, Mathematical Statistics and Probability, Mestrado, Mathematical Statistics and Probability at Instituto de Matemática e Estatística - Universidade de São Paulo (IME-USP)
Doctor of Philosophy - PhD, Statistics, Doctor of Philosophy - PhD, Statistics at University of Michigan
UNESP - Universidade Estadual Paulista "Júlio de Mesquita Filho"
University of Tokyo
University of São Paulo
English, Portuguese, Spanish, Japanese