🧬 About Me
Senior Data Scientist and AI Solution Architect with over 20 years of professional experience at the intersection of process optimization, business logic, and high-performance predictive systems. Specialist in developing end-to-end pipelines that transcend the laboratory research environment to operate robustly and scalably in production environments (MLOps).
Unlike purely theoretical approaches, I master the complete data cycle: from massive ingestion of heterogeneous sources and server-side optimization to containerized deployment of complex neural networks, always focusing on eliminating analytical bottlenecks and delivering real financial return on investment.
🛠️ Skills Matrix (Stack 2026)
| Category | Key Technologies |
|---|---|
| AI & Computer Vision |
|
| Data Performance & Big Data |
|
| Forecasting & Time Series |
|
| DevOps, Cloud & MLOps |
|
🏗️ Strategic Projects
1. Smart Supply Chain AI: Demand Forecasting Engine & Stochastic Data Generation
Integration of high-performance time series predictive modeling and columnar data engineering applied to supply chains.
- Stack:
- Goal: Mitigate catastrophic inventory failures (stockouts) and optimize inventory turnover in complex supply chains.
- Approach: Replacement of slow statistical approaches with modern pipelines using Polars and predictive modeling with LightGBM + MLForecast, incorporating high-frequency macroeconomic and climate exogenous variables.
- Public Validation: Development and publication of a stochastic engine for generating complex synthetic data, adopted by the community for Big Data testing.
- Result: Theoretical reduction in storage costs and optimization of the supply chain logistics pipeline.
- Links:
2. Personalized Medicine: Deep Learning Diagnosis with Ethical Rigor
Advanced pathology classification in medical images focusing on algorithm interpretability and explainability (XAI).
- Stack:
- Goal: Develop clinical decision support systems with high diagnostic accuracy and complete regulatory transparency.
- Approach: CUDA-accelerated custom CNNs and use of AI Explainability (SHAP and LIME) to map decision regions, removing the "black box" problem and meeting regulatory requirements.
- Result: Validated accuracy above 90% in severe scenarios with complete auditability of decision-making.
- Link:
3. Event Analytics & Management Dashboard: High-Performance Server-Side
Interactive business intelligence for operational pipeline monitoring and high-volume financial auditing.
- Stack:
- Goal: Centralize real-time operational financial KPIs to optimize conversion funnels and mitigate latent costs.
- Approach: Modular microservices architecture with decoupled ETL and advanced in-memory caching strategies (
@st.cache_data), reducing latency and RAM consumption. - Result: Month-over-Month executive view with immediate identification of transactional bottlenecks and early detection of billing anomalies.
- Links:
📊 Community Contributions to Data Science & Open Source
- Synthetic Grocery Supply Chain Dataset (Kaggle): I developed and publicly released a stochastic simulator based on real statistical distributions combined with public climate data. The project acts as a Digital Twin generator for simulating transactional data from corporate supply chain networks, facilitating testing and stress-testing of Big Data architectures without breaking legal privacy compliance.
📬 Let's talk?
If you are looking for a strategic partner with solid business maturity to scale analytical intelligence and AI projects from development directly into production: