š AssortApp: How I Built an AI That Reads Documents and Decides Who Should Care
I built an app that takes a document, figures out who in an organization should read it, summarizes it for them, and flags any risks ā all automatically. Her...
Field Notes
Welcome to my Field Notes - a collection of thoughts, insights, and observations from my professional journey. This collection includes articles previously shared on LinkedIn as well as original pieces published here for the first time, offering my perspectives on a broad spectrum of AI, including Data Science, Gen AI, and Agentic AI.
š§ Build Your Own RAG App Using AWS ā Bedrock, Lambda, and API Gateway
Language models are frozen in time. They do not know what happened last week. RAG fixes that. Here is how I built a simple RAG app using AWS Bedrock, Lambda,...
š Beyond Accuracy: Rethinking How We Evaluate LLMs
Why evaluating large language models requires moving beyond accuracy toward semantic, task-aware, and human-aligned evaluation frameworks.
šøļø LangGraph: The Framework That Makes AI Think Like We Do
Discover how LangGraph empowers AI with human-like iterative reasoning, context awareness, and tool useāmoving beyond linear pipelines to intelligent workflo...
š§ How I Used Multiple LLMs to Evaluate Each Other: A Meta-Judgment Experiment in AI Reasoning
Can AI judge itselfāand its peersāfairly? I ran a meta-experiment where one AI wrote a challenging question, multiple AIs answered it, and another AI ranked ...
š Smart RAG Derivatives: CAG, GraphRAG, LightRAG & Contextual Retrieval
RAG has evolved beyond basic retrieval-generation. Explore four powerful derivativesāCache-Augmented Generation for speed, GraphRAG for semantic traversal, L...
š¬ Prompt Engineering Isnāt Dead ā Itās Just Disguised
Despite claims that āprompt engineering is dead,ā it has simply evolved into invisible layers within modern AI systems. From automated frameworks to hidden o...
šÆ Asking the Right Question: How DeepSeek Helped Me Visualize a Conversational Retrieval System
Clarity is currency in technical systems. When I needed to understand RAG architecture better, I asked DeepSeek to create a comprehensive diagram. The result...
š¤ Making AI Agents Smarter: How RAG and Vector Databases Power Custom Intelligence
To build autonomous agents that can reason, retrieve, and act, prompts alone wonāt take you far. You need the logic behind the prompts. Dive deep into how ag...