Went through the xAI PM loop in early 2026. Short version: it's not like any other PM loop I've done. Sharing the long version because I couldn't find anything specific when I was prepping.
Background on me: senior PM, 7 years, mostly B2B SaaS with one consumer product stint.
What the xAI PM loop looks like:
Phone screen with recruiter (standard, see other posts). Technical phone screen with a PM or engineer: they ask about a product you've shipped, and then they push on the technical decisions. Not "what's your roadmap process" but "how did you decide between approach A and B on the backend, and what did you learn."
Onsite had three rounds: Product sense / strategy. They gave me a prompt about improving Grok's user experience for a specific use case. They wanted a crisp prioritization framework, not a 20-slide deck. I got pushback when my ideas were too incremental. They want to see real swings. Metrics / analytical. Define success for a feature, pick metrics, defend trade-offs. Harder than most PM loops. They expect precision. "Engagement" is not an answer; "DAU / retained cohort D7" is closer but they'll still push you. Leadership / behavioral. Cross-functional conflict, ambiguous roadmap situation, managing up. Standard PM behavioral territory.
What was different: every round eventually came back to AI. Can you talk about LLM behavior, evaluation, product limitations? If you're applying as a PM at an AI-first company and you don't have a mental model for how LLMs work, you're at a disadvantage. I'd at minimum read up on prompt engineering, RLHF at a high level, and what "hallucination" actually means technically.
Offer process seemed similar to engineering: 5-8 business days post-onsite.