UP Cebu Computer Science grad who builds at the intersection of AI and Web3. Most recent work spans production LLM systems (multi-agent orchestration, RAG, an AI manuscript-analysis platform) and full-stack product engineering.

Did my internship at ChatGenie (Techstars '23), where I learned to build complete applications from scratch in 3-day sprints using technologies I'd never touched before. Turns out the best way to learn Rails and Vue.js is under pressure.

Recently built a multi-agent AI system that reduces 6-8 hour research tasks to 1.75 minutes through strategic coordination. Learned that having all agents run with shared guidance beats trying to dynamically route.

Also architected a production RAG system where retrieval quality mattered more than LLM quality, and a distributed browser automation platform that taught me WebSocket + job queues is a powerful combination.

I build at the intersection of AI and Web3 - not because they're trendy, but because they're where interesting problems live. Security is an architectural decision in every system I ship.

Started October 2025 learning Solidity from scratch, by November I was building distributed job processing systems, by December I was orchestrating multi-agent workflows. Google Cybersecurity certified.

const daniel = {
  location: "Philippines (Remote)",
  focus: ["AI & Multi-Agent Systems", "Blockchain & Web3", "Security Architecture"],
  recent: "AI Agent Engineer, Contract (LLM manuscript-analysis platform)",
  approach: "Strategic coordination beats dynamic routing",
};

Built multi-agent orchestration system coordinating 7 specialized agents with LangGraph, reducing 6-8 hour research to 1.75 minutes.

Implemented strategic guidance pattern where all agents run with shared JSON objectives. Strategic coordination beats dynamic routing.

Architected production RAG system with 5 retrieval strategies, achieving 67.7% → 85%+ accuracy through hybrid search and cross-encoder reranking. Fine-tuned DistilBERT emotion classifier achieving 92.45% accuracy for MAGSEL research (presented at WILLS 2025, Kyoto).

What I learned:

• Retrieval quality matters more than LLM quality in RAG systems
• Token counting accuracy with tiktoken revealed 72% undercounting in some agents
• Parallel execution with memory-safe state merging patterns apply everywhere

Stack: LangGraph, LangChain, FastAPI, Qdrant, Redis, PyTorch, Hugging Face Transformers

Started learning Solidity from scratch in October 2025. By month's end, shipped full-stack NFT marketplace with ERC-721 contracts, batch minting, and royalty management.

Built blockchain explorer achieving 8-10x performance through LRU caching, and Web3 game with TypeScript SDK abstracting contract complexity.

Comprehensive testing: 52 contract tests plus 50+ integration tests. Not just another tutorial project.

What I learned:

• Gas optimization is an art
• Smart contract testing isn't optional - production systems need comprehensive coverage
• Good abstraction means hiding complexity without sacrificing control

Stack: Solidity, Hardhat, ethers.js, OpenZeppelin, IPFS, WebSocket, TypeScript, PostgreSQL

Google Cybersecurity certified. Security isn't a feature you add later, it's an architectural decision.

Implemented enterprise patterns across projects: Row-Level Security (RLS) for multi-tenant isolation, rate limiting (100 req/15min), input sanitization against injection attacks, secure credential storage, 2-tier fallbacks for reliability.

What I learned:

• RLS at database level beats application-layer checks
• Rate limiting protects both your quotas and your users
• Free-tier constraints force production patterns: graceful degradation, quota protection, resilience by default

Stack: PostgreSQL RLS, Rate Limiting, Input Sanitization, Wireshark, Splunk, Burp Suite, Metasploit

Production ML classification pipeline with end-to-end lifecycle management (Mar 2026)

Code

Fine-tuned ModernBERT-base on Banking77 (77 classes) achieving 91.3% accuracy, 4.52x inference speedup with ONNX INT8 quantization. Automated drift detection with Evidently, A/B testing, feedback loops, and full CI/CD with GitHub Actions. Dagster orchestration, MLflow experiment tracking, Prometheus + Grafana monitoring, Docker Compose (5 services).

200x faster research through strategic coordination

Live Demo | Code

Seven specialized agents orchestrated by LangGraph, turning 6-8 hour manual research into 1.75 minutes. Strategic guidance pattern with JSON objectives, parallel execution delivering 30% speedup, Human-in-the-Loop quality gates, Redis caching protecting API quotas.

67.7% to 85%+ accuracy through advanced retrieval

Live Demo | Code

Production RAG with 5 retrieval strategies: Hybrid Search, HyDE, Multi-Query, Parent-Child, Cross-Encoder Reranking. Redis caching (100x speedup), 2-tier LLM fallback, 100% reliability across 19 test queries. Migrated Chroma → Qdrant Cloud for 2x performance.

Distributed browser automation with sub-100ms monitoring

Live Demo | Code

Production browser automation platform with 23 step types, visual drag-and-drop workflow builder, distributed job processing with BullMQ/Redis, sub-100ms WebSocket latency. Enterprise security (RLS, rate limiting, input sanitization).

52 contract plus 50+ integration tests for production confidence

Code

Full-stack NFT marketplace with ERC-721 smart contracts, batch minting (up to 20 NFTs), royalty management, real-time WebSocket trading feeds, IPFS metadata. Built in October 2025 while learning Solidity from scratch.

More projects: Blockchain Explorer (8-10x performance via LRU cache), MiniWorld (Web3 game with TypeScript SDK), MAGSEL (92.45% DistilBERT accuracy, WILLS 2025 Kyoto)

PyTorch, Hugging Face Transformers, ONNX Runtime, MLflow, Dagster, Evidently, SetFit, Streamlit, Pandera, DVC, scikit-learn

LangGraph, LangChain, Multi-Agent Systems, RAG Architecture, Vector Databases (Qdrant, Chroma), Hybrid Search, Cross-Encoder Reranking, Human-in-the-Loop (HITL), Prompt Engineering, Model Fine-Tuning, FastAPI

Solidity, Hardhat, ethers.js, OpenZeppelin, IPFS, Smart Contract Testing, Gas Optimization, Event Indexing

PostgreSQL RLS, Rate Limiting, Input Sanitization, Secure Authentication, Wireshark, Splunk, Burp Suite, Metasploit

Frontend: React, Next.js, Vue.js, TypeScript, JavaScript, Vite, Tailwind CSS, HTML5, CSS3

Backend: Node.js, Fastify, Express, Django, Django REST Framework, Python, Ruby on Rails, Java

BullMQ, Redis, WebSocket, Socket.IO, Real-time Monitoring, Job Queue Architecture, Caching Strategies

Docker, Git, Vercel, Render, AWS, Linux, Supabase, PostgreSQL, MongoDB, SQLite, Firebase, Cloudflare R2, AWS S3

Playwright, Puppeteer, Workflow Orchestration, Distributed Execution

Mobile (React Native, Flutter), Game Development (Unity, Godot), Algorithm Visualization

Rapid Learning

Proven track record learning under pressure: Rails and Vue.js in 3-day cycles at ChatGenie internship, then Solidity and Web3 in October 2025, followed by Playwright, BullMQ, and distributed systems in November 2025, then LangGraph and multi-agent orchestration in December 2025.

Building something interesting? Want to talk about multi-agent orchestration, smart contract architecture, or how to actually deploy on free-tier infrastructure without compromising quality?

I'm shipping code from the Philippines and open to remote opportunities.