🧬 About Me
AI specialist focused on the intersection of Big Data (Hadoop/Spark) and Applied AI. With a postgraduate degree in Computer Vision, I design end-to-end systems that transcend research environments to operate in production (MLOps).
Moving beyond purely statistical approaches, I master the full lifecycle: from massive "dirty data" ingestion to containerized deployment with Docker/Podman.
🛠️ Skill Matrix (2026 Stack)
| Category | Key Technologies |
|---|---|
| AI & Vision | YOLO, OpenCV, TensorFlow, Keras, Scikit-learn, LightGBM |
| Big Data | Apache Spark, Hadoop, HBase, Flume, Spark Streaming |
| Engineering/Dev | Python, Docker, Podman, DVC, PDM |
| Analysis/Stats | Prophet, Regressions, ANOVA, SHAP/LIME |
🏗️ Strategic Projects
1. Smart Supply Chain AI: Inventory Optimization
Integrating Computer Vision and Demand Forecasting for logistics.
- Objective: Eliminate out-of-stock scenarios and optimize working capital.
- Engineering: Modularized pipeline (YOLO/OpenCV) integrated with Prophet.
- Result: 15% reduction in operational costs.
2. Digital Twin: Synthetic Data Generator
High-fidelity engine for supply chain simulations.
- Edge: Feature engineering correlating weather severity and demand.
- Methodology: Time Series Decomposition using INMET data.
- Result: Massive datasets (100k+ rows) in Parquet format.
3. Personalized Medicine: Deep Learning Diagnosis
Classification of genetic mutations for oncological treatment.
- Technique: Deep Neural Networks (Keras/TensorFlow) with AI Explainability.
- Rigor: Implementation of SHAP/LIME for clinical interpretability.
- Result: Over 90% accuracy in screening.
4. Event Analytics & Management Dashboard
Interactive BI for event management and financial performance.
- Stack: Streamlit, Plotly, and modular ETL architecture.
- Optimization: In-memory processing with caching for large volumes.
- Result: Real-time MoM (Month-over-Month) executive insights.