I build production voice AI systems — specifically STT→LLM→TTS pipelines that handle real customer conversations at scale. As a Forward Deployed Engineer, I've owned end-to-end voice deployments for enterprise clients including Tata 1mg, Dot & Key, and KreditBee, taking automation rates from 77% to 92% and cutting per-turn latency from ~1,600ms to ~1,000ms. My work sits at the intersection of infrastructure and product: I debug root-cause issues across the full pipeline (not just prompt-patch symptoms), optimize for latency and reliability under real traffic, and translate messy production edge cases into concrete engineering fixes. I've driven tool-call success rates to 95% and identified STT failure patterns that fed directly into vendor product roadmaps. Stack: Go, Python, FastAPI, AWS/GCP, SIP/VoIP/CPaaS integration, and LLM orchestration. If you're building or scaling a voice AI product and need someone who can own the pipeline from architecture to production debugging, let's talk.…I build production voice AI systems — specifically STT→LLM→TTS pipelines that handle real customer conversations at scale. As a Forward Deployed Engineer, I've owned end-to-end voice deployments for enterprise clients including Tata 1mg, Dot & Key, and KreditBee, taking automation rates from 77% to 92% and cutting per-turn latency from ~1,600ms to ~1,000ms. My work sits at the intersection of infrastructure and product: I debug root-cause issues across the full pipeline (not just prompt-patch symptoms), optimize for latency and reliability under real traffic, and translate messy production edge cases into concrete engineering fixes. I've driven tool-call success rates to 95% and identified STT failure patterns that fed directly into vendor product roadmaps. Stack: Go, Python, FastAPI, AWS/GCP, SIP/VoIP/CPaaS integration, and LLM orchestration. If you're building or scaling a voice AI product and need someone who can own the pipeline from architecture to production debugging, let's talk.WWWW…