Main activities:

• Support database development, statistical analysis, and AI applications; write up research results; and travel to Brazil to represent UW in meetings with stakeholders.

Deliveries:

• Automation of multiple processes in database development, reducing the time of manual intervention.
• An ML classifier to assess data quality at scale, allowing for the first-time scalable quality check.
• Enhance a deep Learning model to count cattle in protected areas using high-resolution imagery. In collaboration with Google, this project empowers Brazilian prosecutors to act against non-compliant farmers, thereby reducing Amazon’s deforestation.

I work developing end-to-end AI SaaS products to our internal customers.

Main activities:

• Building predictive models using either statistical or machine learning approaches to assess the health of the company assets (tools)
• Assessing the technical feasibility of new projects through data analysis

Deliveries:

• Statistical models to assess the asset health
• Machine learning models to predict the asset health for the upcoming usage
• A custom-trained OCR model to extract dimensions of engineering drawings

Quick facts:

• As result of my first project, our work has been accepted to be published as a scientific paper at OnePetro.
• I led a innovation proposal using AI, and we ranked in the top 12 out of almost 400 ideas worldwide. My colleague and I gave the final pitch to the CEOs.
• I work in one of the most diverse teams in our company, interacting with people from US, India, Europe, and South America in my daily routine.

Tools: Dataiku, GCP, SQL, Python, Dash, Streamlit, machine learning libraries (e.g., scikit-learn), data visualization libraries.

I have worked in multiple roles: facilitator, mentor, and consultant/instructor.

As a consultant/instructor, I prepared the course modules and recorded classes. I recorded four modules: statistics, data cleaning/wrangling, clustering, and model deployment. I have also devised data science activities about descriptive and inferential statistics, unsupervised machine learning models, MLFlow, big data with PySpark, among others.

As a mentor, I assessed student reports and addressed their questions through Q&A sessions, aiding in academic and real-world projects.

As a facilitator, I participated in data science activities from exploratory analysis to model deployment. I also prepared some of those activities.

Squad: Revenue Management

Activities / Deliverables:

• Rule-based automation of pipelines for price engines using PySpark (big data)
• Statistical model to identify price changepoints for several SKU categories
• Monthly time series modeling from of selling volume using state-of-the-art forecast and hierarchical reconciliation methods
• Training for data scientists

Tools: • Python • PySpark • MLFlow • Scikit-learn • Pycaret • Statsmodels • Forecasting frameworks (Prophet, NeuralProphet, StatsForecast)

Activities:

• Advanced statistical modeling: data exploratory analysis, hypothesis testing, time series forecasting, predictive analyses (e.g., regression and classification), customer and product segmentation, etc.

Deliverables:

• Multiple time series forecasting for customer segments (weekly and monthly)
• Model for predicting the probability of the customer recurrence
• In-depth study of churn analysis through inferential statistics

Tools: • Python • PySpark • MLFlow • Scikit-learn • Pycaret • Catboost • Prophet • AWS • GitHub • Data visualization libraries (Plotly, Seaborn, Matplotlib)

Activities:

• Queries and reports for worldwide company members

Grade: 10

• In‑depth study of machine learning models.
• Developed an end‑to‑end hybrid ML model for churn prediction.
• Project code repository

Thesis: Identifying natural selection in Native American populations. Supported by: CAPES (2016 - 2018) and FAPESP (2018 - 2020)

Activities:

• Data analysis, visualization, and scientific reporting of genetic data using R, Python, and bash scripting.
• Application of non‑supervised algorithms (e.g., PCA), descriptive and inferential statistics.

Deliverables:

• Internal R package with customized functions to facilitate multiple analyses
• Three scientific papers published in international magazines
• Thesis code repository
   

I presented our preliminar work in international conferences, including USA. I also took a internship of 3 months in Barcelona - Spain.

Dissertation: Role of cellular prion protein and its ligand, stip1, in the adult neurogenesis. Supported by: CNPq (2014 - 2016)

Main techniques: primary cell culture, immunofluorescence, and hypothesis testing.

Monography: Role of the interaction between the cellular prion protein and its ligand, STI1, in the biology of neural precursors from the murine adult brain. Supported by: University scholarship (2011 - 2012) and FAPESP (2012 - 2013)

Honor & Awards:

• Best academic performance
• Professor Bertha Lange de Morretes’ Award