Vitor Hadad
Applied Scientist II at Amazon Lab126
Menlo Park, California, United States
Summary
🤩
Rockstar🎓
Top SchoolVitor Hadad is an Applied Scientist II with nine years of experience at the intersection of econometrics, causal inference, and adaptive experimentation, currently based at Amazon Lab126 in Menlo Park. He blends rigorous theoretical work—on policy evaluation and valid inference for data collected via contextual bandits—with applied projects that drove measurable social impact, such as increasing contraceptive uptake with the World Bank and improving municipal revenue collection with NYC. A core developer on the widely used grf generalized random forests library, he brings production-grade C++ and R contributions (including optimized sampling routines) to complement his Python experimentation platforms. Vitor has a track record of leading multidisciplinary teams and translating complex statistical tools into practical systems and tutorials for diverse audiences. His background in economics (PhD) and international relations gives him a rare mix of policy intuition and technical depth, focused on thoughtful, welfare-oriented deployment of ML.
9 years of coding experience
4 years of employment as a software developer
University of TokyoDoctor of Philosophy - PhD, Economics, Doctor of Philosophy - PhD, Economics at Boston CollegeMaster of Arts - MA, Economics, Master of Arts - MA, Economics at Simon Fraser University
Github Skills (10)
algorithm10
data-structures10
algorithms10
c-language10
sampling10
cprogramming-language10
random-forest10
sampler10
data-structure10
statistics9
Programming languages (6)
Github contributions (5)
grf-labs/grf
Jul 2018 - Jul 2020
Generalized Random Forests
Role in this project:
Back-end Developer
Back-end DeveloperContributions:19 reviews, 29 commits, 40 PRs in 2 years
Contributions summary:Vitor primarily focused on enhancing the `RandomSampler` class within the `grf` project. Their work included refactoring sampling algorithms, specifically implementing and optimizing the Fisher-Yates shuffle algorithm. They also addressed performance considerations by switching between sampling methods based on the number of samples requested. Moreover, the user made improvements to the codebase to be more consistent with the ranger algorithm.
random-forestdata-sciencemachine-learningforestscausal-inference
halflearned/FHHPS
Aug 2017 - Nov 2019
Linear panel data with two random coefficients
Contributions:136 commits, 4 PRs, 116 pushes in 2 years 3 months
panel-datacoefficientsrandom-coefficientslinearpanel
Find and Hire Top DevelopersWe’ve analyzed the programming source code of over 60 million software developers on GitHub and scored them by 50,000 skills. Sign-up on Prog,AI to search for software developers.