Raj B.

I build production AI agents and LLM systems, including multi-agent architectures, MCP servers, RAG and document-grounded retrieval, and AI customer service platforms across automotive, legal, healthcare, and retail. 3+ years shipping full-stack and AI-driven products. I run my own development agency, Mahant Solutions, taking projects from prompt design and agent architecture through to the database and backend they run on. Selected Work: - AI Customer Service Platform (Automotive): Built the AI layer for a multi-channel customer service agent that reads incoming messages, retrieves facts from business documents and live systems, and replies in-channel. It handles vehicle inquiries, parts booking, and service appointments, with human handoff when a conversation needs a person. Also designed the database the agent queries against. - ECF Email Orchestration (Legal Tech): Built an AI pipeline that parses incoming bankruptcy court notices, including email bodies and attached filings, and runs case-specific analysis across them, such as deriving step payments from a Chapter 13 plan and its confirming order. Extended this with an MCP server so users can query and analyze their own case data using any LLM, not just a fixed dashboard. - Autobox (Multi-Agent Email System): Built a multi-agent system that connects to a user's mailbox and other platforms through MCP connectors, then drafts email replies automatically. It uses the user's past email history for tone and precedent, combined with live context pulled through connected MCP tools. - Medkart Pharmacy (Healthcare/Retail): Worked as a Software Engineer building and optimizing a Warehouse Management System for generic medicine procurement across warehouses in India using Laravel, PostgreSQL, and Redis, while mentoring junior developers.