Dharmendra Yadav — Engineer & Builder

Distributed Systems · AI Engineering · Reliability

This is my portfolio, live at dharmendrra.github.io.

I build large-scale distributed messaging platforms — and, hands-on, AI agents and RAG systems with LLMs, agentic loops (Reasoning + Acting), MCP, and vector databases.

What's inside

• About & Experience — engineering leadership and platform work at Brevo (formerly Sendinblue) and Intecons Software Lab.
• Highlights — re-architecting email at scale, contact-level Kafka fan-out, AI-native engineering, and team leadership.
• Skills — Golang, Python, Kafka, Kubernetes, and an AI engineering stack spanning Agentic AI, RAG, LangChain/LangGraph, MCP, and vector DBs.
• Projects
  • Agentic AI (Python) — FastAPI agent/retrieval/MCP services with conversational memory, PDF retrieval over Pinecone, and live web search with citations.
  • Omni-RAG — multi-backend RAG across ChromaDB, Qdrant, and Pinecone with keyword-sensitive filtering and semantic fallback.
  • Agentic AI (Go) — a Go agent using a Reasoning + Acting loop, tool-calling, and MCP, backed by Tavily search and Ollama-served local models.
  • Rejected.AI — a living lab for early-stage AI ideas and experiments.
• Writing — posts on building RAG pipelines (ChromaDB, Qdrant, Pinecone, MongoDB), an MCP server for MongoDB, and a Reasoning + Acting agent in Go.

How it's built

A deliberately simple, no-build static site:

• Hand-written HTML, CSS, and vanilla JavaScript — no framework, no bundler.
• Blog posts are Markdown files in blog/posts/, rendered client-side and indexed by posts.json.
• scripts/stamp-assets.sh content-hashes CSS/JS into ?v=<hash> query strings for cache busting (wired into a pre-commit hook).
• Hosted on GitHub Pages (.nojekyll, no build step).

Run it locally by serving the repo root with any static server, e.g. python3 -m http.server.

Get in touch

• 📧 [email protected]
• 💼 LinkedIn
• 🐙 GitHub