Solo-built production AI voice-agent platform - 20 yrs telecom/CPaaS/BFSI behind it

I am Rajan Sharma, founder and sole builder of DigitalStaf - a production AI voice-agent platform I architected and shipped entirely on my own.

The system runs a full real-time voice stack: telephony media streaming, speech-to-text and text-to-speech, an LLM reasoning layer, RAG-based retrieval, and an orchestration layer that ties them together. My thesis is simple: in voice AI, orchestration is the product - the hard part isn't any single model, it's making the whole pipeline reliable, grounded and production-ready.

Behind it sits 20 years across telecom, CPaaS, CCaaS, BPO and BFSI - so the architecture reflects how these systems actually behave at scale, not just a prototype.

The architecture is open and public: github.com/digitalstaf/TurnZero

I would like to meet: AI and applied-engineering teams working on voice or agentic systems; enterprises exploring conversational automation who might want a design partner; and anyone building in this space who values someone who can take a system from architecture to running in production, alone.