Voice AI Engineer & AI Agent Builder.
Client Work
Real projects built for real companies — from YC-backed startups to early-stage founders. Each one shipped fast and production-ready.
Freya Voice AI Agent Console
Production-grade voice AI agent console with real-time bidirectional audio streaming. Built STT → LLM → TTS pipeline using Groq Whisper, Llama 3.1, and Cartesia TTS over LiveKit. Includes prompt management with versioning, session analytics, and Dockerized deployment.
OAuth Integration Platform
Built OAuth 2.0 integrations for HubSpot, Airtable, and Notion with a FastAPI backend and React frontend. Implemented token exchange, Redis-backed session storage, and a data loader to fetch CRM contacts and companies post-auth.
Pebble Voice AI Assistant
End-to-end voice AI assistant for collecting business funding information through natural conversation. Full pipeline: mic input → Whisper STT → Groq LLM → Cartesia TTS, with structured data extraction, dynamic document checklist generation, and session memory.
Gensail Real-Time Voice AI
Low-latency real-time voice assistant built with Pipecat over WebSocket. Full pipeline orchestration: NVIDIA Nemotron STT → Groq Llama 3.1 LLM → Kokoro TTS (local). Features Silero VAD for turn-taking, interruption support, and Protobuf-framed audio streaming.
About Me
I’m Sanjay — a Voice AI Engineer and AI Agent builder specialising in real-time conversational AI systems. Shortlisted for Entrepreneurs First Bangalore Fall 2026 (~1–3% acceptance rate).
I’ve built production-grade voice AI pipelines at 133ms latency for Freya Voice (YC S25) and delivered AI systems for UK fintech startup Pebble and VectorShift (YC S23) — full-stack, Dockerized, and shipped in days, not months.
Stack: LiveKit, Pipecat, LangChain, Groq, Cartesia, Whisper, NVIDIA Nemotron, Next.js, FastAPI, Docker, Supabase.
What I Do Best
- ✓ Real-time Voice AI Pipelines
- ✓ AI Agents & LLM Integration
- ✓ LiveKit · Pipecat · LangChain
- ✓ Groq · Cartesia · Whisper
- ✓ Next.js · FastAPI · Docker
Featured Projects
What I Can Help With
Whether you're launching a product or hacking together a prototype, I can help turn your idea into something real and reliable.
AI-Powered MVPs
I build and ship full-stack AI SaaS tools — fast, functional, and tailored for early traction.
Custom AI Integrations
LangChain, OpenAI, Ollama, RAG, vector DBs — I can embed intelligence into your product stack.
Full-Stack Engineering
Frontend, backend, APIs, DBs — I work across the stack with Next.js, Supabase, Prisma & more.
Docker & DevOps
From Dockerization to CI/CD & deploys — I can set up clean infra that scales and survives production.
SaaS Features Setup
Stripe billing, auth, roles, dashboards, emails — I wire it all up so you can go to market faster.
Tech Consulting & OSS Collabs
Need help architecting a system or want to jam on OSS? I’m happy to brainstorm or build together.
What I Can Build
Based on your needs, here's a breakdown of what I can offer. Each project is customizable.
Portfolio & Personal Site
- Responsive site
- SEO + Analytics
- Basic design; +Three.js optional
- Privacy + TOS pages
- Deploy to Vercel/Netlify
AI-Powered Tool
- OpenAI/Ollama/Gemini API
- Chatbot, summarizer, or RAG
- Backend (Node/Flask/Golang)
- Docker + Deploy
Full-Stack SaaS App
- Next.js or MERN
- Auth, DBs, dashboards
- Stripe billing
- CI/CD + DevOps
MVP / Micro-SaaS
- Idea to deployed MVP
- Auth, pricing, frontend, backend
- Minimal UI + optional CLI/API
- Custom AI integrations (LangChain, groq, Gemini)
Automation/Scraping/CLI
- Web scrapers (Puppeteer)
- Automation scripts
- Custom CLI tools in Go/Python
DevOps & Infra Setup
- Docker & Docker Compose
- Nginx / Reverse Proxy
- CI/CD on Vercel / Render
- Basic monitoring & logs