Practical articles.
Grounded in what I've shipped.
Claude Code, MCP, and agentic workflows — explained from the inside. Not surveys or hot takes; write-ups of patterns that held up in production.
Evals and Golden Sets: Notes From Building My First Eval Harness
Why evals exist, what a golden set actually is, the run-grade-decide loop, and a real labels.md file that ranks 14 of my own LinkedIn posts to calibrate an LLM judge. Written while building it, not after.
Read article →
The Testability Seam: Making State-Gated UI Testable by AI Agents
A build-time-gated URL hook that lets AI agents seed chrome.storage.local and reach any state-dependent UI component — plus the production rule that makes it safe.
Read article →
Production-Safe AI Code: A 5-Check Framework Before You Ship
A practical framework for engineering teams shipping AI-generated code. Five checks (correctness, scope, security, performance, observability) before a single AI-written line reaches main.
Read article →
Claude Code Training in India: A Practical Guide in Hinglish
A practical, Hinglish-friendly guide to Claude Code for Indian engineering teams. Installation, daily workflows, Plan Mode, CLAUDE.md, multi-project setups, and Git Worktree, all from production use.
Read article →
How to Install Claude Code on Windows: Complete WSL Setup Guide 2025
Learn how to install Claude Code on Windows using WSL. Complete step-by-step guide covering prerequisites, Node.js setup, authentication, IDE integration, and troubleshooting common Windows issues.
Read article →
Identity Server Jargon & Related Terminology
A complete guide to Identity Server terminology, including authentication, authorization, OAuth 2.0, OpenID Connect, and the differences between Identity Server & IdentityServer4.
Read article →