Production Engineer turned Data Scientist. For 3+ years at Deloitte (Risk Advisory) I turned operational, contractual and financial data into audit evidence, internal-control tests (SOX) and executive reports. Today I build end-to-end analytics: data pipelines, predictive models, exploratory analysis and dashboards that answer a business question.
- 🎓 B.Sc. in Production Engineering, PUCPR (2021) · Postgraduate degree in Data Science, UNIAMÉRICA (2024) · Data Scientist program, DNC School (2026, 292 h)
- 💼 Business Analyst, Risk Advisory @ Deloitte (Sep 2022 – Jan 2026)
- 🧭 How I work: CRISP-DM, from the business question to the model and the dashboard
- 🌐 Languages: Portuguese (native) · English, EF SET 75/100 (C2): Reading 99 · Listening 72 · Speaking 53 (B2)
| Project | What it shows | Stack | |
|---|---|---|---|
| 🏠 | Airbnb Rio: Data Pipeline | End-to-end cleaning and transformation of listings and reviews: missing values, IQR outliers, categorical encoding | Pandas · NumPy · Seaborn |
| 🛒 | RFV Customer Segmentation | RFV scoring and rule-based segmentation of 4,233 e-commerce customers, with an interactive dashboard: Champions are 20% of customers and 61% of revenue | Pandas · Seaborn · Recharts |
| 🧱 | Concrete Strength Prediction | Regression benchmark (Linear, Ridge, Lasso, Decision Tree, Random Forest, Gradient Boosting) against baselines, with feature importance and a Streamlit simulator | scikit-learn · Streamlit |
| 💰 | Data Science Salaries | Statistical analysis of salary trends with an interactive Plotly Dash dashboard | Dash · Plotly · Pandas |
| 🦠 | SafeConnect | COVID-19 incidence forecasting for 183 countries (JHU CSSE, World Bank), validated on an unseen future against baselines; neighbour features from a geographic country network improve the forecasts | scikit-learn · NetworkX |
| 💳 | IEEE-CIS Fraud Detection 🚧 | Fraud detection on the IEEE-CIS dataset: exploratory analysis done, modeling in progress | Pandas · Seaborn |
📚 More projects and study repositories
| Project | What it is |
|---|---|
| Structuring Company's Stock | Stock management system and executive dashboard built in Excel (pivot tables, KPIs) |
| IA Colab Preparation Agent | Desktop utility that turns Markdown lessons into Google Colab notebooks via the Drive API |
| Book Publication Analysis | Exploratory analysis of book publishing data: genres, ratings, sales and publisher revenue |
| DNC Data Science | Data Scientist program, completed in 2026 (292 h): Excel, Power BI, Python, statistics, regression, classification, clustering, PySpark and Streamlit apps |
| DIO · Santander Bootcamp | Santander Bootcamp 2023 study notebooks: Python fundamentals |
| Contact Book · Priority Task Manager | DNC Python cases: interactive command-line apps with JSON persistence |
🔒 Some of my most recent work, LLM-powered applications built with FastAPI, RAG and React, lives in private repositories. I'm happy to walk you through it in an interview.
🐍 Languages & Core
📊 Data Analysis & Visualization
🤖 Machine Learning & AI
🖥️ Apps, APIs & Databases
🧪 Also used in projects and coursework
PySpark · PyCaret · Hugging Face Transformers · Gradio · statsmodels · Google Drive API · Tkinter · PyInstaller · Playwright · pytest
- 🥇 Champion & 3rd Place, Sustainability Hackathon: predictive and business solution focused on efficiency and social-environmental impact
- 🔬 Undergraduate research (PIBEP): applied research on validation and logistics models
- 🏛️ Deloitte Risk Advisory: SOX internal-control testing (TOE), risk mapping and process modeling for a global industrial joint venture
I'm interested in data roles where analytics meets business risk and operations. The best way to reach me is LinkedIn.
"The key isn't just knowing the code or commands, it's understanding what to do with them."
Sérgio Arena