Two rounds total, and the interviewers barely showed any expression the whole time. Two rounds total: the first was with someone pretty senior who asked me about a past project of mine, question-and-answer format, behavioral. The second round was coding plus introducing another one of my projects. They mainly work on conversational AI, and the questions didn't dig into specific technical details — it was pretty general, just repeatedly asking what the data was, what the model was, how the eval worked, and then that was it. The coding question was about counting connected islands after flooding, where islands touching the boundary don't count — a classic DFS problem, medium-to-upper-medium difficulty I'd say. My coding is only average, so I kept double-checking back and forth the whole time, and it took about half an hour.
Honestly I do like the application of agents and LLMs in the medical field — two of my projects are about this, and they got published at major ML/NLP conferences. They said the main reason they'd hire me is that they want me to publish papers. The work itself does seem like a match, but doing pure research also feels kind of strange to me. They said it's because interns rarely have time to actually get data and build something real, so instead you can use open-source data to write some papers first — that seems reasonable enough, but if I'm not touching the actual business, then coming here for an internship doesn't feel like it means much. Mainly given the recent unfavorable news about Oracle Health, if I want this to be a safety net for a return offer, it's probably not a great place to be — who knows if this department will even still exist next year, it's rough. Since I've spent most of my time so far publishing papers and haven't had many business-facing internships, I'd still rather go somewhere where I can be involved with the business and demonstrate real business impact.
Discussion
Loading comments…