hey, i'm rakshit.

i'm a senior software engineer at meta in london, and the short version of what i do is: i work on keeping people's accounts safe at a scale that's honestly hard to wrap your head around sometimes.

most recently i've been leading the push to move facebook beyond passwords entirely. that means shipping passkeys (biometric logins) for facebook web, so that signing in is both safer and less of a hassle. on any given day i'm writing heuristics to catch account compromises before they happen, wiring up integrations like the google risc api, and figuring out how to stay compliant with eu regulations without breaking everything. my stack is hack, python, react, and graphql if you were curious.

i studied information and computer engineering at cambridge, where i got genuinely hooked on the idea of using maths to make sense of messy real-world data. neural signals, financial markets, you name it. that curiosity eventually led me to co-author some research on using temporal cnns to predict short-term price movements in crypto order books, which was equal parts fascinating and humbling.

outside of work, the thing i keep coming back to is reinforcement learning. there's something that really clicks for me about teaching an agent to figure out the world through pure trial and error. it feels like the rawest form of intelligence. i write about rl, algorithms, and system design here on this site (what used to be called "the boring blog").

when i'm not at my desk, i enjoy cricket and keeping active at the gym.