AI Engineer
- Building a multi-tenant AI voice-calling platform: bots with versioned prompts, phone and browser test calls, a call queue with retries and circuit breakers, campaigns, recordings, transcripts and LLM post-call analysis (FastAPI, MongoDB, LiveKit, Docker).
- Made every speech and model vendor a plugin (Sarvam, Deepgram, ElevenLabs, Murf, Fish Audio, Gemini, OpenAI) and benchmarked 12 TTS models, so each bot can switch providers on latency, cost and language quality.
- Shipped multilingual calling agents (English, Hindi, Telugu) with tool calls for demo booking on Google Calendar, scheduled callbacks and hand-off to a human.