Posts

Building Multi-Agent Systems the A2A Way: A Hands-On Project with LangGraph

Image
  A claims processing system with three services that communicate using Google's A2A (Agent-to-Agent) protocol over JSON-RPC 2.0 : 🔹 A Gateway that handles auth, guardrails, and routing 🔹 A Claims Agent that validates claims and checks for fraud 🔹 A Policy Agent that answers coverage and benefits questions Each agent publishes its own capabilities at a standard /.well-known/agent-card.json endpoint — meaning any A2A-compatible client can discover what it does without me hard-coding anything. Same pattern you'd use to plug into a partner agent in real production. 🛠️ The stack ✅ FastAPI + Pydantic for the JSON-RPC layer ✅ LangGraph for stateful agent reasoning (StateGraph + checkpointing) ✅ MCP (Model Context Protocol) for clean tool abstraction ✅ Anthropic Claude — using Haiku for routing, Sonnet for reasoning, Opus reserved for hard cases ✅ LangSmith for end-to-end observability ✅ Custom guardrails for PII detection and prompt injection 🔍 Why this combo clicks M...
Post a Comment
Read more

T20 Cricket Score Prediction using a full RAG pipeline

Image
  I spent the last few weeks building something I've always wanted — a GenAI app that actually understands cricket. Not just "who's winning" — but why , and what to do about it. Here's what I built: T20 Cricket Score Prediction using a full RAG pipeline. The stack: Google Gemini 2.5 Flash as the LLM brain Pinecone as the vector database for semantic search over match stats sentence-transformers for local embeddings (free, no API cost) CricAPI for live T20 data Streamlit for the web UI + Docker for deployment The interesting part? I gave it this prompt: "Analyze the 2026 T20 World Cup Super 8 standings. Calculate the exact win/loss margins India needs to qualify over West Indies. What happens if Sunday's match gets washed out? And what does Pakistan need against Sri Lanka?" It came back with NRR calculations, rain-rule implications, powerplay strategies for the trailing team — all backed by retrieved data, not hallucination.That's RAG working as ...
Post a Comment
Read more

The Internet's Treasure Hunt: How Your Computer Finds Websites

Image
 Have you ever realized that computers don't actually understand website names? When you type  https://something.com  on your laptop or phone, your computer has no idea where that is. Computers only speak in numbers, specifically  IP Addresses  (like  192.0.2.1 ). The system that translates the human name you know into the numerical address the computer needs is called  DNS (Domain Name System) . Look at the diagram below. It might look like a confusing map of arrows, but it’s actually a very organized treasure hunt. Let’s break down the story of how your request travels through this map using an analogy of trying to find a specific room on a giant school campus. Step 1: The Quick Check (The DNS Cache) Look at the stick figure on the left. That’s you trying to go to  something.com . Before your computer goes out to the internet, it does the equivalent of checking its own pockets. It looks in the  DNS Cache  (that first diamond shape). Th...
Post a Comment
Read more

AI's "Brain" is Stuck in the Past. Here's How We Give It a Library Card.

Image
 We’ve all been there. You ask an AI chatbot for help, and it confidently tells you something that's... well, completely out of date. "I'm sorry, my knowledge only goes up to 2023." It’s frustrating. These incredible "brains," known as Large Language Models (LLMs), are like the smartest person you’ve ever met. They’ve read almost the entire internet, but they’re stuck in a "closed-book" exam. They don't know what happened yesterday, and they certainly don't know anything about  your  private company documents or your school's homework list. So, how do we fix this? We can't spend millions of dollars retraining these giant models every single day. The answer is surprisingly clever, and it’s called  RAG (Retrieval-Augmented Generation) . Forget the jargon. At its heart, RAG is just a story of a great partnership. Here is that explanation adapted into a blog post with a human, conversational tone. AI's "Brain" is Stuck in ...
Post a Comment
Read more

Diffie-Hellman Key Exchange (TLS Handshakes)

How Your Browser Secretly Shares a Key (Even on Public Wi-Fi!) Ever see that little padlock 🔒 in your browser and wonder how it  actually  keeps your stuff safe? How can your browser and a website agree on a secret code to scramble your password... when any hacker could be listening in? It sounds impossible, right? It's not magic, but it's a super clever mathematical trick called the  Diffie-Hellman key exchange . At its core, it's a way for two parties (like your browser and a server) to create a  shared secret key  over the public internet, without  ever sending the key itself. This brand-new shared secret is then used to create  another  key, called a  symmetric key  (you might hear tech folks call it 'AES'). This  second  key is a super-fast, heavy-duty encryption key that does all the hard work of scrambling and unscrambling your data for the rest of your visit. The easiest way to get this is with the famous "mixing paint...
Post a Comment
Read more