I build production AI systems: agents, RAG pipelines, MCP servers and the web data that feeds them. Full-stack engineer at Discover WebTech, Indore.
What I work on
- Agents — multi-agent workflows with LangGraph and LangChain, human-in-the-loop, tool use
- RAG — ingestion, chunking, hybrid retrieval and evaluation with LlamaIndex and LangChain
- MCP — remote MCP servers with OAuth, SDKs and agent tooling
- Web data for LLMs — crawlers that turn bot-protected sites into clean markdown
- Full stack — Next.js, React, TypeScript, Python, FastAPI, Postgres, MongoDB
MeshArc — a web crawler and scraper for AI agents: search, scrape, map and crawl sites to clean markdown, and track what changes.
| Repo | What it is |
|---|---|
| mesharc-mcp | MCP server: crawl and scrape websites from any MCP client |
| mesharc-python | Official Python client, with the MCP server built in |
| mesharc-node | Official Node.js client |
| claude-code-plugin | MeshArc for Claude Code: search, scrape, map and crawl from the terminal |
agentic-rag-platform — an open reference app for agentic RAG: LangGraph agent, LlamaIndex retrieval, FastAPI streaming backend and a Next.js chat UI. In active development.
NEXUS — a multi-agent system that analyses calls and meetings across voice, language, face, body, gaze and conversation, then fuses them to catch what a single channel misses.
- punkpeye/awesome-remote-mcp-servers#1879 — add the MeshArc remote MCP server
Python TypeScript LangGraph LangChain LlamaIndex MCP FastAPI Next.js React Node.js Postgres MongoDB Docker