I'm Rishi. I write here.
Most of my work is backend: infrastructure, keeping the product stable, and a fair amount of time on the customer side, reading feedback and digging into what people actually hit when they use what we build.
Lately that has meant LLMs, mostly getting them to read documents without inventing things. Almost none of the hard part turns out to be the model call. It's everything around it: what survives text extraction, where you split the work, and how you find out whether a change made anything better.
This site is where I write that down. Not tutorials and not advice, just what I tried, what broke, and what I'd do differently. I'm working it out as I go and writing it while it's still fresh, mostly so I don't forget.
I'm on X as @rishipatidarr. That's the best way to reach me.