Anthony Tecsa

Built a DAG-based workflow engine with Zustand and Supabase; scaled to 250+ weekly runs across 4 businesses.

Developed real-time LLM agents with WebSocket streaming to OpenAI and Twilio; cut latency by 83%.

Architected secure onboarding flows for organizations, enabling user authentication, role-based access, and encrypted API key management

Built a librarian AI agent using a query-routed RAG pipeline with OpenAI, boosting book recommendation accuracy by 15%.

Achieved 90%+ test coverage with Jest and Postman, cutting production bugs by 30% through early defect detection.