I’m a Data Scientist and Machine Learning Engineer passionate about merging AI, analytics, and user interfaces to create impactful end-to-end products.
From building scalable fraud detection systems to NLP-powered insight engines, I love solving problems where data meets design.
- 🔭 Current: Lead Data Scientist @ AR Data Technologies — building IoT & BIM ML pipelines.
- 🧠 Previously: Eskwelabs Fellow — socio-economic prediction models (92.56% accuracy).
- 🌱 Learning: Generative AI & agentic systems for automated insight extraction.
- 💬 Ask me about: ML pipelines, NLP, Marimo dashboards.
- ⚡ Fun fact: I used to design buildings as an architect — now I design data systems!
Core Skills:
- ML & AI: XGBoost, TensorFlow, PyTorch, Deep Learning, NLP, LLM Concepts
- Data: Pandas, NumPy, SQL, Power BI, Seaborn, Matplotlib
- Web: React, Tailwind CSS, Flask, Streamlit, Cytoscape.js
- Other: Docker (learning), AWS MLOps (in progress)
Cloud-native financial data platform designed to ingest hourly stock market data, compute technical trading signals, monitor portfolio-level risk metrics, and track historical signal performance.
The platform simulates backend infrastructure that could power financial analytics products used by active traders, quantitative analysts, and small investment firms.
AI-powered building code compliance for AEC. This project helps architects and designers check early space planning (rooms, doors, corridors) against building codes and internal standards, without relying on full BIM. It combines Vision LLMs (blueprint extraction), RAG (building code Q&A with citations), and deterministic compliance checking in a single proof-of-concept aimed at future CAD Add-In integration (AutoCAD/Revit).
🧑🏫Canva Slides
End-to-end scalable ML pipeline reducing manual fraud review by 4×.
Built with XGBoost, Cytoscape.js, and React, featuring 20+ node fraud cluster visualization.
🔗 Live Demo
NLP pipeline analyzing AI perception in Philippine media (2020–2025).
Used spaCy, Selenium, and BeautifulSoup to extract sentiment and trends.
Deep Learning model (TensorFlow) predicting user repurchase behavior with 80% accuracy.
Achieved 97.5% accuracy using a two-hidden-layer NN — core deep learning fundamentals.
- 🌐 Building AI-driven dashboards with Streamlit and React
- 🧮 Experimenting with Agentic AI for automated analytics pipelines
- 📊 Developing visual storytelling with Power BI + Python
AR Data Technologies — Lead Data Scientist (2025–Present)
→ Designed data pipeline for IoT & geospatial ML systems.
→ Architected early-stage MLOps dashboard and rule-based prototype.
Eskwelabs — Data Science Fellow (2025)
→ Built Gradient Boosting model (92.56% accuracy) & skill-network analysis using centrality metrics.
VAA Philippines — Amazon PPC Specialist (2023–2025)
→ Automated 40+ performance reports, boosting ad ROI by 20–30%.
