Pranav Pendhari

I build AI agents and automation systems that run in production — not demos that fall apart the first time a real user or a real edge case hits them.
I work across industries: customer support and lead-handling agents, internal workflow automation, document/data-intelligence pipelines, and custom AI features bolted onto existing products. Over 2.5+ years of hands-on work, including inside a regulated financial-services environment, I learned to build agents with proper human checkpoints, error handling, and observability by default — not bolted on after something breaks. I’ve also won a Technology & Innovation award at a national-level agentic AI hackathon, competing against 18+ teams from major global consulting and technology firms.
What I build (any industry):
• Custom AI agents & multi-agent systems (LangGraph, Google ADK) — sales, support, ops, research
• RAG / knowledge-base assistants — chat with your docs, product data, or internal wiki
• Workflow & process automation — turning manual, repetitive review/approval work into agent-driven pipelines
• Document intelligence & data extraction pipelines
• AI feature integration into existing web apps / SaaS products (FastAPI + React)
Specialized depth (premium track):
• Compliance & audit-grade agents for insurance and financial services — human-in-the-loop approval gates, structured audit logs, explainable decisions. This is where I command the highest rates, because most freelancers building agents have never had to satisfy a compliance officer.
Stack: Google ADK, Vertex AI (Gemini), LangGraph, LangChain, RAG/vector search, FastAPI, Python, React/Vite, GCP (Cloud Functions, BigQuery, Document AI), Docker/CI-CD, MCP-pattern tool orchestration.
Whether you need a quick AI-powered feature shipped or a fully auditable agent system for a regulated workflow — let’s talk about what you’re trying to automate.