I'm an AI Engineer and startup Co-Founder who builds intelligent, agentic systems that solve real-world problems at scale.

Currently I'm Co-Founder & CTO of Brrdcast, a hyperlocal AI-powered knowledge network where people broadcast location-based questions and get real-time, AI-curated answers from their neighborhood. I own the platform end to end — from LangGraph multi-agent pipelines and geospatial broadcasting to an always-on WhatsApp conversation engine. Before Brrdcast, I led AI development at ihiring.ai, architecting a production-grade AI voice interview platform that scaled to 500+ concurrent sessions.

What I'm building:

• Brrdcast — a hyperlocal AI knowledge network spanning a web app, an admin dashboard, and an always-on WhatsApp engine. LangGraph supervisor agents (Query + Reply pipelines), hybrid retrieval (Pinecone vector + MongoDB BM25 with RRF), H3 geospatial broadcasting, and SSE streaming.
• Jarvis — a multi-agent "Personal AI Operating System" for founders, built on a Perceive → Understand → Plan → Act → Communicate loop. A hub-and-spoke orchestrator routes to 7 specialized Claude sub-agents, with a TrustEngine approval gate and TriSearch retrieval across Qdrant, Postgres FTS, and Neo4j.
• CHIMERA — an AI-agent red-teaming framework for security-testing LLMs, agentic apps, RAG pipelines, MCP servers, and AI coding assistants against prompt-injection and chain-composition attacks.

My journey:

I've spent 5+ years evolving from web development to AI engineering across startups and enterprises. At Loginsoft I built the Cytellite IP enrichment platform processing 10M+ daily records with 40% latency reduction, and developed healthcare systems serving 50K+ monthly users. Earlier, I built full-stack applications, integrated payment gateways handling $500K+ monthly transactions, and architected cybersecurity research platforms.

What drives me:

• I'm fascinated by the intersection of AI research and production systems. How do you take cutting-edge LLM capabilities and make them reliable, scalable, and useful? That's the puzzle I wake up solving.
• My approach combines solid software engineering (distributed systems, microservices, cloud architecture) with modern agentic AI (multi-agent orchestration, RAG, knowledge graphs, the Model Context Protocol). I believe the best AI systems are built on boring, reliable infrastructure.

Tech stack:

• AI/ML: LangChain, LangGraph, OpenAI, Anthropic Claude, Multi-Agent Systems, Agentic AI, Model Context Protocol (MCP), RAG, Knowledge Graphs, AI Red-Teaming, Transformers, Embeddings, AWS Bedrock
• Backend: Python (FastAPI, Django), Node.js, Go, gRPC
• Frontend: React, Next.js, TypeScript, Tailwind CSS
• Data: PostgreSQL, MongoDB, Redis, Qdrant, Neo4j, Pinecone, Elasticsearch
• Cloud & DevOps: AWS (Lambda, S3, SQS, EventBridge, Bedrock), GCP, Docker, Kubernetes, Terraform

Currently exploring:

LLM orchestration, AI agent frameworks, AI security, and production ML systems. Recent certifications: Machine Learning Specialization (Stanford), LangChain & LangGraph (DeepLearning.AI / LangChain Academy), Agent Communication Protocol, and Claude Code.

Let's connect if you're:

• Working on interesting AI/ML problems
• Building production AI or agentic systems
• Hiring for AI Engineering roles
• Just want to chat about technology

Always happy to help fellow engineers navigate the AI landscape!

• 📧 Email: shivadharmi@gmail.com