Satyam Namdev

ezAIx, Inc · Thousand Oaks, USA (Remote)

• –Architected VoiceCoreIQ backend — 18 Django apps, intelligent routing, queue orchestration, and omnichannel messaging.
• –Built production AI Voicebot with Azure OpenAI GPT-4o Realtime, enabling sub-second latency and tool-calling (order lookup, CSAT).
• –Engineered RAG pipeline: Azure Blob → Azure Document Intelligence & Firecrawl → llama-text-embed-v2 → Pinecone namespaces.
• –Designed node-based visual workflow engine for dynamic call routing, IVR, and omnichannel orchestration.
• –Developed real-time Agent Presence Dashboard via custom ASGI SSE middleware and Redis Pub/Sub for live supervisor monitoring.
• –Architected event-driven Azure integration using Event Grid and webhooks for incoming calls, media streams, and call recordings.
• –Reduced TTS costs by 85% and call setup latency by 70% via 3-tier caching (Redis → Blob Storage → Generation).
• –Flew to Dubai (Dec 17–22, 2025) for the company kickoff — selected engineers and founders gathered in person for a week of strategy and planning.

TunableLabs, LLC · San Francisco, USA (Remote)

• –Built end-to-end Legal-AI system — FastAPI backend, Next.js frontend, Supabase (PostgreSQL) for multi-tenant chat sessions and document collections.
• –Designed LangGraph-based Entity Extraction Agent with schema validation and multi-agent orchestration across 5,000+ legal documents.
• –Implemented multi-LLM orchestration, streaming responses, and real-time WebSocket updates; integrated live web search via Tavily and Exa APIs.
• –Refactored a tightly coupled Gradio monolith into a modular FastAPI backend, enabling scalable API integration across multiple law firms.

Multi-turn streaming RAG chat over 468 curated ideas with multi-provider failover across Groq Llama 3.3 70B and Gemini 2.0 Flash. Built a custom inference logging SDK for latency, token usage, errors, and PII redaction per request, plus an observability dashboard with real-time metrics and Langfuse traces.

Multi-agent system using LangGraph to automate medical case analysis — processing 500+ patient notes, lab reports (LlamaParse), and radiology images (Llama 4 vision) with 94% accuracy. SOAP notes and diagnostic suggestions in under 3 seconds.

Custom-trained YOLOv8s on 5,493 images — 90% detection accuracy. OpenStreetMap API integration to map 1,000+ km of road data for geolocated pothole reporting. Dockerized deployment.