George Gondim

Data Scientist, and Machine Learning engineer with academic and industry experience in software development and Machine Learning/Deep Learning having worked in different areas including machine learning, deep learning, computer vision, distributed systems, front-end, and back-end development.

Nubank | 03/21 | Lead Data Scientist (full-time)

• Currently, the Tech Lead of two Machine Learning models for the Lending Business Unit with director-level stakeholders, accountable for their whole pipeline including conceptualization, development, validation, approval, deployment, and maintenance.
• Conceptualized, developed, deployed, and maintained a new risk machine learning model to predict personal loa risk, reducing losses by ~20% across all Lending products.
• Mentored and onboarded eight new Data Scientists/Machine Learning Engineers by pairing to explain the tech stack, setting up roadmaps, and deliverables, and aligning expectations with their stakeholders.
• Interviewed more than 30 candidates for Data Scientist and Machine Learning Engineers positions.
• Promoted from Senior to Lead Data Scientist.

• Generated around a 4% increase in revenue from 500+ Facebook and 100+ Google ad campaigns by developing a system to cluster and split marketing campaigns.
• Sent ~150 KPIs visualizations to stakeholders daily by developing an Airflow system that automatically produced and sent them via email across the company.
• Designed and analyzed quasi-experiments to measure intervention effects on marketing campaigns.
• Leveraged data to produce actionable insights to drive decisions for War Machines (Android and iOS game), responsible for ~30% of the company’s revenue.
• Promoted from Associate to Journey Data Scientist.

University of Campinas (UNICAMP), Campinas, SP, Brazil, 03/17 – 07/19, Master of Science, Electrical Engineering, GPA 4.0/4.0

• Relevant course work: Neural Networks I and II, Stochastic Processes, and Evolutionary Computing.
• Orally presented paper “Adversarial Attacks on Variational Autoencoders” – Gondim-Ribeiro et al., 2018 at the 1st
• Official Latinx in AI Research Workshop, Conference on Neural Information Processing Systems (NeurIPS) in Montréal, Canada (2018) (acceptance rate 21%).
• Awarded scholarship from the São Paulo Research Foundation (FAPESP, Brazil).