I build AI-native developer tools, TypeScript libraries, and full-stack product prototypes.
My current focus is the space where practical software engineering meets AI systems: LLM orchestration, security utilities, reactive runtimes, realtime applications, and product-shaped prototypes that can grow beyond a demo.
Selected work, live products, and the broader story behind what I am building.
| Project | Focus | Why it matters |
|---|---|---|
| FluxDOM | Compiler/runtime experiment | Custom .flow compiler, direct DOM generation, Vite integration, monorepo architecture |
| llm-conductor | AI orchestration | Multi-provider LLM flows, Zod-backed structured outputs, TypeScript library design |
| omniguard | Security toolkit | Validation, sanitization, hashing, JWT helpers, native Web Crypto |
| NimbleJS | Reactivity | Signals, computed values, effects, stores, persistence, framework-agnostic state |
| BuddsApp | Realtime product | Supabase chat, LiveKit voice rooms, stories, roles, AI companion behavior |
| Solar Breakdown Predict System | Operational dashboard | Solar-field monitoring, FastAPI backend, QR technician workflow, ML-oriented artifacts |
- TypeScript, JavaScript, React, Vite
- Node.js, Express, Python, FastAPI
- Supabase, PostgreSQL, SQLite, realtime systems
- LLM application patterns, structured outputs, prompt workflow design
- Web Crypto, validation, sanitization, JWT utilities
- Compiler/runtime experiments and frontend architecture
I am especially interested in:
- AI tooling and agentic product workflows
- startup-style full-stack engineering
- developer experience and small libraries
- systems that can move from prototype to production
- GitHub: onuracar-dev
- Email: onuracar.work@gmail.com
