Personal details
Bhargav T.
About
7+ years building production AI systems across healthcare, enterprise SaaS, and conversational AI, specializing in low-latency voice platforms, LLM applications, Retrieval-Augmented Generation (RAG), and realtime communications.
Founding AI engineer at MEGA.AI, where I helped build an enterprise Voice AI platform from the ground up, designing Python services powering AI voice agents over Twilio, Vonage, LiveKit, and WebRTC. Experienced building production speech pipelines combining Deepgram, Cartesia, Gemma, OpenAI, Anthropic, Azure AI, and custom LLM deployments while optimizing latency, reliability, and operational cost.
Strong background across backend engineering, AI infrastructure, React Native applications, cloud deployment, and production observability.
Work Experience
Founding AI engineer responsible for designing and scaling production Voice AI systems for outbound debt collection and AI-powered customer support.
- Designed and implemented low-latency Python services powering AI voice agents across Twilio, Vonage, WebRTC, and LiveKit.
- Built production speech pipelines combining Deepgram speech recognition, Gemma LLMs, Cartesia speech synthesis, and turn detection models for natural voice conversations.
- Developed browser-based LiveKit communication workflows supporting AI-powered customer service sessions with realtime audio streaming.
- Integrated multiple AI providers including OpenAI, Anthropic, Azure AI, Deepgram, ElevenLabs, and Cartesia through a modular provider architecture.
- Optimized end-to-end voice latency across WebSocket connections, speech recognition, LLM inference, and text-to-speech generation while reducing infrastructure costs through custom model hosting and audio caching.
- Migrated production inference infrastructure from Cerebrium to Azure while maintaining service reliability.
- Designed automated AI-to-AI simulation tools to evaluate conversational behavior before production deployment.
- Implemented production monitoring and observability using Grafana and Sentry to accelerate issue diagnosis and improve system reliability.
Skills & Technologies
Python • LiveKit • WebRTC • Twilio • Vonage • Deepgram • Cartesia • Gemma • OpenAI • Anthropic • Azure AI • WebSockets • Pipecat • Grafana • Sentry • Azure • Cerebrium • vLLM • Ollama • Docker
- Built AI systems for KompetencyAI, a competency assessment platform for healthcare and pharmaceutical organizations.
- Contributed to multi-agent conversational systems simulating clients, colleagues, and supervisors for realistic role-play assessments.
- Developed configurable evaluation pipelines, automated grading, analytics dashboards, and governance workflows.
Skills
Python • LangChain • Multi-Agent Systems • LLMs • Healthcare AI • Analytics