Available for AI engineering work
Engineeringintelligentexperiences.
I'm Sharad, a Generative AI Engineer & Vibe Coder at The BAAP Company. I design and deploy real-time voice agents, low-latency LLM backends, and automated agentic workflows using FastAPI, LangChain, and Gemini.
voice_agent_daemon.py
LIVE
RTSP Feed 30 FPS
Vector DB Qdrant
Model Latency 184ms
Audio stream input
3+
Years building AI
40%
Operator load reduced
<200ms
Realtime chat latency
99.9%
System reliability
Scroll to explore
What I do
Focused on real-time, production AI.
Voice AI
Real-time STT → LLM → TTS pipelines with sub-second turn taking.
Agentic Workflows
Composable LangChain agents with planners, tools and retries.
LLM Backends
FastAPI + WebSocket services tuned for low latency and reliability.
Production Debugging
Owning live AI systems — observability, eval loops, hot patches.
Selected work