About

I’m Anupa Adikary, a machine-learning engineer working in computer vision. I hold an MEng from Oxford, where I specialised in machine learning.

My work centres on perception systems — getting models from a promising result on a benchmark to something that runs reliably in the real world. I’m especially interested in vision transformers, efficiency at scale, and the engineering that sits between a paper and a deployed model.

I write here to think in public. My aim is to do frontier research without a PhD — so this site is the evidence: the ideas I’m chewing on and the things I’ve built.

Reach me at anupa.adikary@gmail.com or on GitHub.