1st round: initial screening 2nd round: take home challenge which involved building a chat bot which can answer questions based on financial documents. The final result is assessed on LLM accuracy. 3rd round: some LLM theory warm up questions + AI production Systems design architecture drawing and live code challenge for LLM orchestration

Initial 30 minute call with another solution engineer, and was then given a take-home task to complete, with a follow-up call with one of the founders. The task was very under-specified, both in terms of the desired functionality and the aspects of the solution they want to focus on (if you're going to assess things like automated testing, CI/CD, code format, or whatever, on a 4 hour task, it might help to let people know this beforehand so they know which corners it is ok to cut). I spent quite a few hours over a weekend on the task, and was ready to discuss the pros and cons of the solution, but I didn't get asked a single question about it during the call. I had the strong impression that no-one at Tomoro even looked at my solution (which is borne out by other reviews here). Instead, I was asked questions about the design of a solution he'd recently worked on (always amazed by people who supposedly have a data or stats background but who will happily assess candidate suitability on a random sample of 1 question).