Senior Data Scientist

• Marketing Mix Modelling planning and execution for leading enterprises of all industries in Brazil, using Bayesian Inference algorithms in a fully automated and scalable environment using Airflow and MLflow.
• Alpha tester for Google's Meridian Bayesian MMM framework, providing detailed feedback and reporting on Meridian's implementation and performance on real-world marketing data and comparing to other MMM implementations.
• Research and Implementation of a Feature Interaction framework to allow for deeper understanding of complex interactions in non-linear models, enriching MMM's interpretability and providing better insights.
• Development and maintenance of Machine Learning modelling tools and frameworks to increase productivity and enable fast and robust model execution, using Pandas, PyStan, SKLearn and Plotly.
• Refactoring of old deployment code for enabling automated model execution and retraining, reducing complexity and time to value to production model.
• Planning and management of end to end Machine Learning projects, including Business Model discovery and design, data request structuring, model hypothesis planning and validation, model testing documentation and results presentation, using Asana and Google Workspace.

Data Scientist - Store Demand, Supply and Assortment

• Creation of an end-to-end pipeline for Demand Forecasting using LightGBM, Dask, MLflow and Airflow, reducing traning time by 50%, costs by 90%, and enabling model lifecycle and performance tracking.
• Development and maintenance of the store supply tool using PySpark, Pandas, PuLP and Airflow, responsible for optimizing distribution of dozens of thousands of products daily across over 1400 stores.
• Creation and maintenance of dozens of data ETL pipelines using PySpark, Delta Lake, Airflow and BigQuery.
• Creation and maintenance of queries and dashboards for metrics and KPI tracking as well as making data more accessible to the whole company, using Google BigQuery and Looker Studio

• Development and deploy of a Facebook integrated Chatbot based on sentence similarity language models, using AWS Lambda, API Gateway, DynamoDB and Docker.
• Data mining and pre-processing of structure and non-structured data (text and images) using Python, for model training and analysis.
• Development of an object detection and recognition model using transfer learning and deployment using AWS Lambda, S3 and API Gateway.
• Time series momentum algorithm implementation for quantitative trading of financial assets, with a backtesting library, using Pandas and Numpy.