Writing
Building with LLMs, AI for analytics, and keeping good software affordable with open source. Published here first — and sometimes mirrored to Medium.
Most "AI" still just talks. This is how I built two agents that act — one that turns plain English into structured Jira stories, one that branches, commits, opens a GitLab MR and reviews it — using LangChain, local models, and one surprisingly small loop.
A beginner-friendly, step-by-step walkthrough of putting a website live on AWS for about a dollar a month — buying the domain on GoDaddy, then wiring up storage, HTTPS, a CDN, and DNS, with the real gotchas nobody warns you about.
After 200+ articles on Medium, I'm moving my writing home — here's the thinking, and what changes for you.
A look at how AI Mitra uses open-source building blocks plus an LLM layer to stay capable and affordable.
Part 7 of my Beginners Guide to Building AI Agents — one agent, 15 tools, from branch creation to merge with AI code review baked in. The architecture, the design decisions, and how to adapt it.
Part 6 of my Beginners Guide to Building AI Agents — giving an agent write access. You type one sentence; it creates a complete, well-structured Jira story (summary, acceptance criteria, points, epic, labels) in seconds.
Part 4 of my Beginners Guide to Building AI Agents, and the first with a real-world use case: a ~150-line Python agent that reads my work notes (or my Jira tickets) and writes my daily standup for me.
Six cognitive-psychology principles, eight weeks of engineering, one mission — make Indian students love learning again. Inside the AI Mitra alpha launch.
A first-hand account of running an open-source analytics stack on bare metal — and the honest numbers when we priced out Databricks, Snowflake, and Fabric.