I'm Daniel Ruffolo, an AI Product Engineer with 8 years in IT, focused on cloud platforms, automation, and AI agents. I build systems that take the repetitive work off people's plates β from healthcare applications where accuracy is critical, to self-hosted infrastructure I use as a testing ground for new ideas.
- WardBeat: a clinical-operations copilot for hospital ward flow. Cited discharge barriers, a grounded copilot and next actions a human approves. Next.js + FastAPI + NVIDIA NIM + pgvector.
- AI Patient Flow Orchestrator: a multi-agent perceive β reason β plan β act loop over a simulated hospital, with a human approval gate on every action.
- itsaagent-cli: a local-first ReAct agent CLI on Ollama. No cloud, no API keys.
- trello-mcp-service: a Trello MCP service with local Ollama and Telegram integration.
- personal-planner-agent: a personal-planner agent in ~300 lines of plain Python (Ollama + SQLite). Companion code for a tutorial.
- ai-resources: agents and skills for Claude and OpenCode.
- nextjs-fullstack-boilerplate: Next.js 16, Auth.js v5, Drizzle + Postgres, PWA, Docker, CI.
- nextjs-rag-boilerplate: chat with your PDFs, with cited answers. Built on the template above.
- cloudflare-hono-d1-boilerplate: React on Pages, Hono on Workers, D1, Firebase Auth.
- react-express-firebase-boilerplate: React + Express monorepo with Firebase Auth and Postgres.
- rabbitmq-tutorial: work-queue patterns with RabbitMQ, Python and Docker.
- chatbot: a Telegram Q&A and reminders bot on a local LLM.



