i'm going through the xAI process for a data/infra eng role and the info online is thin. most threads i find are a year old or talking about grok product feedback, not the actual interview.
if anyone has done their loop recently (2025 or early 2026), drop what you remember: how many rounds, what types, what the technical focus was, how fast they moved. role type matters too since MRE and SWE are probably different loops.
appreciate any signals, even vague ones. this community usually has the freshest data.
5 replies
infra_ines
did an infra/platform screen about 5 weeks ago. one round technical, focused on distributed systems and fault tolerance. they gave me a scenario: your training job crashes at 95%, what do you do. less about the right answer and more about how you reason through it and what questions you ask.
staff_steph
went through a staff SWE loop late last year. 5 rounds total, one was a values/culture conversation that was surprisingly substantive. not the usual "tell me about a conflict." more like a genuine discussion about how you think AI development should go. be prepared to have actual opinions.
de_derek
that's useful. did they push back on your opinions or were they just taking notes? i'm wondering if contrarian views on AI timelines would be a red flag there.
staff_steph
they pushed back a little, in a curious way not a confrontational way. i think genuine uncertainty is fine. overconfident doomism or hype probably both read badly. just be honest about what you actually think.
sec_sasha
worth noting that the process probably varies more than people think at a company this size. different hiring managers run different loops. one person's experience isn't a template.