Part of the core team at Vox Radar, a social media monitoring start-up, offering a SaaS product used byclients to monitor real-time topics such as politics, sports, marketing campaign outcomes, and more across various social media platforms and news outlets. As a Senior Data Scientist, I am one of the leaders in the data science division, overseeing the use of allstatistical and machine learning algorithms within the company. These algorithms are used for a varietyof purposes, including natural language processing (NLP), face and object detection in images and videos,sentiment analysis, applications of recent AI technologies for summarizing and classifying posts, as well as statistical modeling to identify potential disruptive growth of certain topics in real-time. I have contributed to projects for clients such as Lojas Americanas, CNN, Ambev, and Energisa.Vox Radar has received considerable media attention in Brazil for its comprehensive data analysis studies, published in renowned newspapers.On a recent project, we collaborated with a prominent financial asset management firm in Brazil to leverage Twitter and news data, significantly improving on the models previously available to forecast Brazilian inflation. Our advanced methodologies were documented in a working paper called "What do news and social media tell us about future inflation?", of which I was a co-author. I presented our findings at both the "Fourth Brazilian Meeting of Data Science," hosted by FGV University, São Paulo, and the "Fifth International Workshop in Financial Econometrics."

At Kogno, we specialize in combining data analytics with economic and financial insights to offer bespoke solutions spanning data science, econometrics, and machine learning. Our portfolio includes predictive modelling, pricing strategies, and corporate solutions. As apartner, I worked on several consulting projects, some of which are described below.*** Lojas Americanas Project (Jul' 17 – Feb ’23): Partnered with a leading PUC-Rio lab and one of Brazil's premier retail giants, Lojas Americanas. The project revolved around using mathematical modeling to forecast sales for various department items and assist in pricing strategies and promotional offers definition, using mainly R and Python.Key Contributions:- Developed a web-based application designed to support decision-making within the company. Thistool allowed employees to define desired pricing parameters, while remaining factors were determinedthrough an optimization process, considering variables such as stock levels, days of the week, price,and various other characteristics identified by the model.- Collaborated with the team to create models to forecast demand and estimate elasticity of retailproducts, delivering data-driven insights and recommendations, which played a crucial role in shapingretail and stock strategies and promotions.- I presented a poster detailing some of the project’s methodologies and outcomes at the "FourthInternational Workshop in Financial Econometrics."*** Itaú Asset and Itaú BBA Training Courses (Feb ’21 – Dec ’21): Conducted specialized training for Brazil’s largest private bank, focusing on practical applications of statistics and machine learning in financial forecasting.*** Duty Cosmetics Pricing Consultancy (Sep ’21 – Dec ’21): Advised Duty Cosmetics on pricing strategies for their Brazilian market debut.

Key contributor to the creation of covid19analytics.com.br, a pivotal platform in Brazil's COVID-19 response. Our platform forecasted virus trends and contamination rates at city, state, and national levels, garnering extensive media coverage with hundreds of interviews and mentions throughout Brazilian media.

Taught an undergraduate course in R programming within the Department of Economics. The courseemphasized practical business applications, including forecasting, data handling and cleaning, andbasic modeling techniques.

Worked in risk and pricing, where I was responsible for the daily distribution of risk and managementreports. Enhanced the department’s routine risk operations using Excel and programming skills, andachieved an integration between R and Excel using .NET programming.

Optimization of the Likelihood Function of GAS (Generalized Autoregressive Score) Models via Classical Methods and Heuristic Methods: application to Wind Series from Northeastern Brazil.