Rahul

Python, FastAPI, WebSockets, Pipecat, Twilio, Gemini (Live API), Deepgram, ElevenLabs, Sarvam, real-time streaming, STT/TTS pipelines, LLM integration, Docker

Pushing Voice AI latency below 500ms, multi-agent voice architectures, and voice observability/evals for production phone agents.

Writing the "Production Voice AI" series — deep dives into latency, barge-in handling, and why voice agents fail after the demo. Building AI phone agents that survive real calls.

Voice AI / AI phone agent engineering roles (remote, international), consulting on real-time voice pipelines, and technical discussions about production Voice AI problems.