Aasim Sani

MeowLearning

See on Github

This is a pick-your-problem style guide I created to educate everyone, from my leadership team to my ML engineers on working with AI in production settings. This is the stuff you won't learn in most ML/AI courses.

Artificial Manga Panel Dataset

See on Github

I love manga, but I can't read Japanese, and Google Translate's OCR + translation pipeline isn't great with manga text. So I set out to build a speech-bubble detector to localize manga into English faster (which would also help official translators). There was no free, publicly available dataset to train it on, so I generated one.

Second Brain Slipbox

Sneak a peek of my second brain

I read a lot and listen to a lot of podcasts, and most of it leaks out within a week. So I built a Second Brain using the Slipbox Method (Zettelkasten): a system that organizes what I want to keep into an interconnected web.

gwt: Git Worktree Manager

See on Github · PowerShell port

Stupidly simple git worktree management for ZSH (and PowerShell). Smart branch naming, interactive pruning, and automatic config copying so you can hop between parallel workstreams without losing your shell setup. Built because I kept losing track of half-finished branches and re-cloning repos to escape merge-conflict purgatory.

ML Project Template

See on Github

An open-source template repository to kickstart new Machine Learning projects in a standardized way. Originally built for CopyCat's ML team (saving ~2 hours per project), it works for any ML team that wants a repeatable starting point.

Agile Connects

Energy savings algorithm development

An algorithm to find the upper bound on energy and cost savings when AgileConnects' IoT scheduling decides when AC units cycle on and off.