I'm Kaleem Molani.

What I actually do

By day I'm a Tech Lead at a Noida-based software company. Most of my work lives in the unglamorous, important middle of things: connecting CRMs, moving companies off systems they've outgrown, wiring up integrations, and lately building AI agents that do useful work instead of just talking. The hard part is never the code — it's understanding the mess well enough to design something that survives contact with reality.

Why I do this in the open

Honestly? Because I learn better when I have to explain myself. Writing something down forces me to actually understand it — and the half-finished, slightly-wrong version is usually more useful to someone a few steps behind me than a polished tutorial written years later. So I publish the learning as it happens: the wins, the dead ends, and the things I got embarrassingly wrong before they clicked.

How I think

A few things I believe, that show up in everything I build:

• Systems before tactics. Most problems handed to me as a "quick fix" are system-design problems in disguise. Fix the system and the symptoms stop coming back.
• Model reality first. Get the underlying model right before automating anything — name things correctly, map how they actually relate. Automation on top of a wrong model just makes the wrong thing happen faster.
• Clarity is a deliverable. If I can't explain a design simply, I don't understand it yet. The spec, the diagram, the plain-English version — that's not overhead, it's the work.

The rest of me

When I'm not doing that, I tinker with microcontrollers — ESP32s and the like — building small hardware gadgets for the pure fun of making something physical. I'm building an AI curriculum on the side. And I teach: my own team, and a couple of people one-on-one. I've come to think explaining things is just how my brain processes them — this site is the public version of that habit. I'm based in Jodhpur, in Rajasthan, India.