🌐 https://bertolucci-rl.github.io/
Welcome to my portfolio!
Here you’ll find a selection of projects focused on Data Science, Machine Learning, and AI Engineering, with an emphasis on building practical, production-ready systems.
I’m a Data Scientist and AI Engineer with a Master's degree in Mathematics (Functional Analysis/Operator Theory) from UNESP.
My work focuses on building practical, production-ready systems using machine learning and data-driven approaches. I combine a strong mathematical foundation with hands-on engineering to develop reliable, real-world AI solutions.
- Machine Learning & Data Science
- AI Engineering & Backend Systems
- Predictive Modeling
- Data-driven decision systems
End-to-end ML system for clinical risk prediction with API and UI, achieving ~0.92 AUC and high recall for disease detection.
Stack: PyTorch, FastAPI, Streamlit, SHAP
Impact: Designed as a decision-support tool for rapid triage, prioritizing recall to minimize undetected cases.
🔗 https://github.com/ric-rky/heart-disease-predictor
Interactive Streamlit application for exploring and comparing ML algorithms through visualizations of decision boundaries, metrics, and learning dynamics.
Stack: Streamlit, scikit-learn, PyTorch, Plotly
Impact: Improves intuition and model interpretability through real-time experimentation and visualization.
🔗 https://github.com/ric-rky/interactive_ml_playground
End-to-end quantitative pipeline including deep learning (feedforward NN, LSTM), anomaly detection, risk metrics (VaR/CVaR), and Monte Carlo simulation.
Stack: PyTorch, pandas, scikit-learn, SHAP
Impact: Combines predictive modeling and risk analysis to simulate future scenarios and support financial decision-making.
🔗 https://github.com/ric-rky/finance_ds_project
Regression system with feature engineering, log-transformed targets, and model comparison, achieving strong performance (R² ≈ 0.99 with tree-based models).
Stack: scikit-learn, pandas, statsmodels, Streamlit
Impact: Supports real estate pricing decisions with accurate predictions and an interactive interface for end users.
🔗 https://github.com/ric-rky/modelo_pred_alugueis_completo
Predictive and decision-support system for collections, including probability estimation, Next Best Action (NBA), and cost-aware optimization.
Stack: scikit-learn, XGBoost, LightGBM, SHAP, Streamlit
Impact: Enables data-driven collection strategies, optimizing actions based on expected return and reducing unnecessary interventions.
🔗 https://github.com/ric-rky/case_inadimplencia
- AI Engineering and system design
- LLMs and applied AI systems
- Model evaluation and reliability
- Backend integration for ML systems
- GitHub: https://github.com/ric-rky
- LinkedIn: https://www.linkedin.com/in/ricardo-bertolucci/