# Marc Doria — Technical Portfolio

> Undergraduate Computer Science student at University of the Philippines Manila specializing in Applied AI, Explainable Machine Learning (XAI), and Full-Stack Systems. National Champion and Global Finalist in L'Oréal Brandstorm 2026. Published author in IEEE IISA 2026.

## Core Information
- Website: https://marc-d.is-a.dev/
- GitHub: https://github.com/MarcD25
- LinkedIn: https://linkedin.com/in/marc-doria
- Full Agent Context: https://marc-d.is-a.dev/llms-full.txt
- Resume: https://marc-d.is-a.dev/res
- Curriculum Vitae: https://marc-d.is-a.dev/cv

## Featured Projects & Research
- [ClaimCheck](https://marc-d.is-a.dev/#projects): Multimodal Generative Engine Optimization (GEO) and Vision-Language Model attribution engine with Grad-CAM perturbation attribution (Undergraduate Thesis). Tech stack: Python, PyTorch, Gradio, HuggingFace, Grad-CAM.
- [Decoding SLE](https://marc-d.is-a.dev/#projects): Pathway-grouped explainable machine learning for Lupus risk prediction using TreeSHAP and GWAS Catalog, achieving 0.812 AUC-ROC (IEEE IISA 2026 Full Paper). Tech stack: Python, XGBoost, TreeSHAP, Scikit-learn, Reactome.
- [MyNaga Health Suite](https://mynagahealthsuite.netlify.app/): Civic digital health suite with digital IDs, voice-to-form AI medical triage, and guardian adherence tracking for 155,000+ residents (1st Naga City Mayoral Hackathon Grand Winner). Tech stack: React, Node.js, Groq, HuggingFace, Netlify, PWA.
- [GJeep](https://gjeep-livedemo.web.app/): Modernized public transit payment and tracking system with real-time telemetry and contactless BLE fare collection (GCash ImaGnation Top 4 Finalist). Tech stack: React, Firebase, TypeScript, BLE, IoT Sandbox.
- [Jot-it](https://jot-itdemo.netlify.app/): Full-stack intelligent task management system with natural language task creation, persistent contextual assistants, and JWT authentication. Tech stack: React, TypeScript, Node.js, Supabase, PostgreSQL, Cloudflare Workers, Tailwind CSS.
- [Project: Tanda](https://marc-d.is-a.dev/#projects): Automated pipeline converting complex PDF documents into semantic Anki flashcards via OCR and multimodal embeddings. Tech stack: Python, FastAPI, PaddleOCR, ChromaDB, Ollama, CLIP.

## Key Competencies
- Programming Languages: Python, TypeScript, JavaScript, C, C++, SQL, LaTeX, HTML/CSS, PHP
- AI, ML & Data: PyTorch, Scikit-Learn, XGBoost, TreeSHAP, HuggingFace, FastAPI, Power BI
- Web, Cloud & Tools: React, Node.js, Express.js, Supabase, Firebase, Cloudflare Workers, Vite, Tailwind CSS, Docker, Git, PostgreSQL