An evals engine for people who ship AI features
A small, opinionated harness for testing LLM features against real user tasks instead of leaderboard benchmarks.
I'm Rohit, a product manager who can't stop making things on the side. This is where I publish what I'm building, what I'm writing and what I'm listening to, as it happens.
A small, opinionated harness for testing LLM features against real user tasks instead of leaderboard benchmarks.
A small electric two-wheeler rental business I run on the side: fleet ops, pricing and a lot of WhatsApp.
A personal publication with its own CMS, built with Next.js and Supabase. You're reading it.
Small, scripted environments for seeing where agents go off the rails, and how they recover.
An experiment in turning practice sessions into short loops you can actually stick with.
Fifteen minutes I come back to every November. No commentary needed.
Benchmarks tell you how a model does on someone else's problem. Your users only care about theirs.
Two scooters, one spreadsheet and the fastest feedback loop I've ever had.
Four days, one motorcycle, a lot of fog. Photographs from the road.
Every tooltip tour I've shipped was an apology for a default we didn't have the nerve to choose. Pick the default. Let people change it later.
If a chart doesn't change what someone does on Tuesday, it's decoration. Most dashboards are a gallery.
The clever version needs data you'll only get by shipping the boring one.
I build products for a living and build other things for fun. Most weeks that means one serious project, a few half-finished experiments, and a playlist I can't stop editing.
Product manager, builder, occasional musician.