Local LLM Visual Bench
Ask local models to draw cherry blossoms and solar systems. Compare the HTML artifacts side by side. No leaderboard scores — look at the output.
Hi, I'm Eeshan.
Human, artist, technologist. I lead data science at Starbucks. In my free time, I experiment with the latest in AI & data science, while also making time for my passions in art, poetry and writing. Hope you enjoy exploring my site.
Technical Writing
all 13 →Aug 9, 2026Local AI Series #6: minimal-ai: a simple tool to run local models as coding agents on a MacNewJun 19, 2026Please don't let AI write for youJun 14, 2026Local AI Series #5: Apple is talking about it too nowMay 29, 2026Local AI Series #4: Testing local AI models by asking them to analyze a real A/B testMay 16, 2026Local AI Series #3: Testing local AI models by asking them to build Cherry Blossoms and Solar SystemsMay 3, 2026Local AI Series #2: The Regular Person's Guide to Running AI on Your LaptopApr 26, 2026Local AI Series #1: It's time for the regular person to start testing local AI modelsApr 6, 2026How I De-Vibed a Vibe-Coded NLP App
Personal Blog
all 150 →Oct 10, 2026You think you’re having a bad day?NewSep 15, 2026Pop Culture Philosophy Series #1: Linkin ParkSep 8, 2026Book Review: Lolita by Vladimir NabokovSep 6, 2026Oil Pastel #2: Dream Lake HouseSep 3, 2026New art medium unlocked: Oil PastelsAug 29, 2026Sketching is easyAug 28, 2026Watercolor is hardAug 19, 2026Ambivalence
Projects
all →How I Prompt
Wrapped-style analytics for your AI coding conversations. Discover your prompting persona, see metrics, get a shareable report. Everything stays local.
Sidequests
Track every side project you've ever started. Scans your dev folder, surfaces health, GitHub state, and AI-powered next actions. One command to install.
A/B Testing Memory Game
A live experiment you can play. Real users, real data. The analysis found a significant effect — but completion rate dropped from 89% to 69%. Don't ship it.