Data Scientist

• Designed and developed the data ETL pipeline to compute massive financial datasets, reducing computation time from several days to hours (PySpark - Databricks - Azure);
• Created a ”methodology on code” to guide new Machine Learning projects development, from which two projects were structured;
• Developed Machine Learning Models for forecasting Bitcoin Trading Volume (XGBoost, 65% accuracy) and Crypto Volatility (Temporal Fusion Transformer, 95% accuracy);
• Developed an MLOps pipeline to put models to production: calculation of hundreds of features, streaming inferences in real-time using FastAPI Websocket server, model monitoring using SQL database and Slack API (Docker on ECS, AWS FeatureStore, FastAPI, SQL, Slack), with an uptime of over 99%;
• Created a numerical Index to score Time Series data and help build trading strategies (statistical correlations, co-integration, ACE score, time series analysis).

Performed customer churn statistical analysis for a large Brazilian B2C video streaming app.

• Developed algorithms for GPS filtering, correction, and performance analysis for amateur football players, including lite football athletes.
• Took charge of ensuring the quality of data products and managing tasks such as backlog, OKRs, and metrics.
• Integrated Python algorithms with AWS on the Serverless Framework, including Lambda, S3, EC2, DynamoDB, and CloudWatch.
• Carried out tasks such as data analysis, implementing Kalman Filters, Neural Networks, Fuzzy Filters, scientific algorithms, statistical insights engine, basic computer vision, and facial recognition algorithms.
• Utilized Python to create powerful data visualizations.
• Maintained the Python infrastructure for the automatic delivery of hundreds of daily.