Giuliano Oliveira

Built an in-house daily demand forecasting and safety-stock model for newly introduced items, solving the cold-start problem and replacing the incumbent Neogrid vendor solution. Cut stockouts from 10% to 6.4% with an average increase of only 2 items in stock.

Built the first data infrastructure on GCP from scratch for a retail company with no prior data culture. Designed and deployed an AI-driven CRM qualification system over WhatsApp using direct API calls with tool calling.

Served as Agile Lead on a GenAI project for Scania, owning the roadmap and client communication while supporting development of an agent system on Amazon Bedrock. Built and deployed end-to-end ML models for inventory forecasting for Embraer using BigQuery, BQML, Python, SQL, and Terraform.

Developed ML models and statistical analyses for Ambev in Python and Databricks, quantifying the return on leased point-of-sale equipment. Built an ETL pipeline and a mixed-effects model forecasting a competitor's sales volume at 5% average error.

Led forecasting and classification projects for clients such as JBS, Leroy Merlin, and Engie, applying time series, regression, and ML on big-data stacks. Developed a stock loss prediction model for Leroy Merlin and contributed to data-driven products for Azure Marketplace.

Pursuing a Bachelor of Science in Systems Analysis and Development, expected to graduate in January 2028.

Completed the Apache Spark Programming certification from Databricks.

Completed the Data Science Complete Bootcamp on Udemy.