Founding AI Engineer · Voice AI · Forward Deployed
Building real-time AI systems that work at 3AM, not just in demos.
I specialize in production voice AI and LLM-powered systems — the kind that handle thousands of calls a day and can't afford to break.
- 🎙️ Now: Contributing to pipecat-ai/pipecat (10.5k ⭐) — automatic failover, telephony fixes, pipeline reliability
- 🔬 Research: Unsupervised detection of hesitation in speech and video — no labels, real signals
- 📍 San Francisco
Python Pipecat WebRTC LangChain RAG HuggingFace FastAPI Spark Docker
- Hardening Pipecat: A Month of Fixing What Matters — dev.to
- The 2AM Call That Voice AI Developers Dread — Medium
Most of my recent work and open source contributions live at: github.com/kollaikal-rupesh