I assumed leadership of the data science team in the demand planning area during the migration of predictive AI models to a new proprietary planning platform. I led this transition with a focus on stability, governance, and alignment between technology and business.

Under my leadership, we developed predictive AI models for forecasting demand across new brands and sales channels within the group, reaching the highest number of production forecasting models ever recorded in the history of Grupo Boticário. This achievement was enabled by building a highly technical team, advancing our MLOps pipelines, and establishing a development flow that accelerated incremental deliveries without compromising robustness, scalability, or business alignment. I also led experiments to continuously improve existing models, significantly impacting metrics such as WMAPE and Bias, while fostering a culture of solution longevity and scalability through our maturity in software engineering and MLOps.

I led initiatives to increase planners’ adherence to the models, such as the development of a forecast reliability indicator and an explainability module that allows users to understand the impact of factors such as seasonality, cannibalization, price elasticity, post-promotion dip, and promotions on the predictions. I also led a strategic project in partnership with an external S&OP consultancy to test new predictive approaches, in which our models were validated as market benchmarks.

With a focus on diversity, I led hiring efforts that resulted in a gender-balanced team (50% women). I foster a collaborative, technically rigorous, and positive work environment, with a genuine commitment to the well-being and development of the team. As a result, we maintain high engagement scores on the Team Culture platform, with highlights in Leadership, Happiness, Well-being, Feedback, and Personal Growth.

As a senior data scientist in demand planning, I worked during the structuring of the corporate data platform on Google Cloud Platform (GCP), contributing to the validation of dozens of tables and translating critical logistics network rules into a structured and democratized table within the new architecture.

During this period, I worked closely with the business, applying approaches such as Dynamic Time Warping in an A/B test to assess changes in franchisees’ purchasing behavior due to a proposed change in the pending order cancellation process. The analysis resulted in a solution proposal, which was successfully implemented.

I was also responsible for rewriting a legacy demand forecasting model in Python using LightGBM, ensuring its integration with the new demand planning platform developed by Grupo Boticário’s technology team.

I joined the Data Science team at Grupo Boticário with the objective of developing predictive solutions to support the demand planning team in decision-making. I was responsible for developing a predictive AI model based on Gradient Boosting to forecast the probability that a statistical forecast would require adjustment by demand planners. This solution became essential in helping teams prioritize which forecasts to adjust, in a context of overload and high behavioral uncertainty in sales channels during the COVID-19 pandemic.

In addition to modeling, I was responsible for the entire deployment, monitoring, and maintenance process of the solution, ensuring its operation in a production environment with governance and traceability.

I worked on the analytics team, collaborating directly with the customer relationship area. I developed the company’s first churn probability model, which resulted in this scientific article.

I used natural language processing techniques to analyze consumer reviews and comments, proposing strategic actions to improve service quality and, consequently, enhance the ratings of the company’s official store on marketplaces such as Mercado Livre and Americanas.

I regularly attended Multivariate Analysis and other statistics courses taught by Professor Anselmo Chaves from UFPR, held on company premises. I also participated in workshops and academic events, including the 2019 Operational Research and Logistics Symposium held by the Brazilian Navy in Rio de Janeiro.

I was responsible for developing and delivering an 8-hour Google Sheets training for employees across various departments, in addition to contributing to the development of the new onboarding process for new hires.

During a career transition, I was hired by Hilab as an intern at the age of 32, while pursuing a Bachelor’s degree in Mathematics as a second graduation. After four months, I was promoted to Data Analyst.

I contributed to the implementation of Power BI in the company, working from the design of the data warehouse on Amazon Redshift and the development of data ingestion pipelines on Amazon S3, sourced from Azure Cosmos DB, to the creation of interactive dashboards and reports for various departments.

Additionally, I participated in a Growth Hacking committee composed of members from different areas, focusing on evaluating the company’s main challenges and levers, raising and prioritizing hypotheses, and modeling experiments.