Tiago is an experienced Data Scientist and Machine Learning Engineer fascinated by how data, predictive modeling, algorithms, and statistical methods may be used to solve complicated industry challenges. With a varied academic background in political science, Computer Science, and AI, he strives to incorporate ideas and information from multiple disciplines wherever he works. Tiago has hands-on experience as part of the Data Science/AI team at Banco do Brasil, where he developed NLP algorithms for multi-class classification challenges.
Worked along the Campaign Budget Optimization team at Pinterest to refactor their CBO (Campaign Budget Optimization) product, improving code readability, developer experience, and maintenance.
Analyzed data regarding ads budget, assisting announcers to improve their performance and expends per ad.
Researched and implemented relevant ML algorithms and tools.
Developed NLP classification models using Python, TensorFlow, MLOps, TensorFlow Extended (TFX), and TF Serving.
Worked in a private cloud environment (Docker and Kubernetes) on a project designed to process about 30 million requests per month 24/7 and is integrated with the Brazilian OpenBanking platform.
Developed features that assisted clients to organize their expenditures and the company to design better products.
Assisted Banco do Brasil, one of Brazil's largest banks, in designing, training, deploying, and integrating an online text categorization model for customer spending with older systems, providing millions of customers with a better financial solution and a better insight into their spending. Used Python for data analysis, data annotation, model training, etc., TensorFlow Serving, JavaScript/Node.js for the integration API, and Docker for CI/CD deployment in a private cloud.
This project used mathematical models to assist advertisers in lowering their budgets and increasing the turnover of their Pinterest campaigns. Assisted in the development of a simulator for marketers' expenses and in reworking and refining the tests and testability. The majority of this project was created in Java.
Built, improved, and maintained the Personal Bank website for Banco do Brasil clients, who can use this personal bank system to obtain statements, credit card information, and loan information, among other things. Worked with Java, JavaScript, HTML, and CSS.
Education
Data Science - Basics
Data Science Academy - Brazil
2020 - 2020
Artificial Intelligence Specialist
IESB - Brazil
2018 - 2019 (1 year)
Degree in Computer Science
UniCEUB (Centro Universitário de Brasília) - Brazil