Thiago Martins
Cofundador at Stealth Startup
Rio de Janeiro, Brazil
Summary
🤩
Rockstar🎓
Top SchoolThiago Martins is a data-driven CFO and former VP of Data with 11 years of industry experience and a 15+ year background spanning finance, analytics and actuarial science. Based in Rio de Janeiro, he built and scaled the data team at Rei do Pitaco, led integrations between finance and data functions, and drove initiatives that improved bottom-line performance by 70% and generated R$5M in procurement savings. A PhD-trained statistician and former Principal Data Scientist at Yahoo, he created PyVespa and contributed to Optuna-LightGBM integrations, showing a rare blend of production ML engineering and financial leadership. He excels at translating complex models into actionable financial projections, automated controls and fast month-end closes. Adjunct professor at FGV, he combines academic rigor with hands-on product and platform development across AI, pricing, experimentation and risk systems. This mix of deep quantitative expertise and operational finance makes him adept at turning advanced analytics into measurable business impact.
11 years of coding experience
10 years of employment as a software developer
Pós-graduação Lato Sensu - Especialização, Finance and Financial Management Services, Pós-graduação Lato Sensu - Especialização, Finance and Financial Management Services at COPPEAD UFRJM.Sc, Statistics, M.Sc, Statistics at Universidade Federal do Rio de JaneiroNorwegian University of Science and Technology
Portuguese, English
Github Skills (8)
hyperparameter-optimization10
machine-learning10
pytest10
python10
integrations10
lightgbm10
scikit-learn9
scikit9
Programming languages (8)
Github contributions (5)
optuna/optuna
Aug 2020 - Aug 2020
A hyperparameter optimization framework
Role in this project:
ML Engineer
ML EngineerContributions:2 reviews, 6 commits, 1 PR in 1 day
Contributions summary:Thiago focused on enhancing the integration of the Optuna hyperparameter optimization framework with the LightGBM library. Their contributions involved adding and refining metric aliases, ensuring the correct metric names were used within the tuner and associated tests, and updating test cases to incorporate these changes. These modifications aimed to improve the usability and flexibility of the framework by allowing users to define metrics using different names.
hyperparameter-optimizationpythonmachine-learningparalleldistributed
Escavador/pyvespa
Sep 2020 - Nov 2022
Contributions:1135 commits in 2 years 2 months
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.