Further AI is a Series A startup that automates insurance underwriting. I interviewed for Staff Applied AI Engineer. The whole process was 3 rounds.
Round 1:
HM phone screen (15 minutes), mainly to see if there's a fit. We talked about projects I had done before, and they asked about my experience with and interest in using AI. Pretty relaxed overall.
Round 2:
Two 45-minute coding sessions.
Merge Intervals. A very simple classic, no follow-up.
Build a service with AI. The requirement was a service that creates and executes workflows, with two APIs: create workflow and execute workflow.
What the interviewer really wants to see is how you use AI, not how much code you end up writing. They care about: Plan first, then Implement, then Verify correctness.
My advice: first work through the requirements and design together with the AI, then start writing, and in the end make sure you add tests and actually run them to verify. Don't just hand everything to the AI and hope for the best.
Round 3:
Onsite (4 sessions in total)
Project Deep Dive. You talk through a project you've done, with the focus on how you used AI in it. You need to cover:
What problem you were solving, what was technically hard, what architecture decisions you made, what alternatives you considered, the trade-offs and lessons learned, and the final business or customer impact.
My advice is to prepare a project with a complete storyline ahead of time, especially the alternatives and trade-offs part, since they dig into that quite a bit.
Product Intuition. This tests how pragmatically you work with a product team: cutting scope when you need to, solving the customer's most important problem first, and only then thinking about extending features. When you answer, give plenty of real examples of "shrinking scope and shipping a minimum viable version first".
CTO interview. More of a chat: your interests, why you're looking, and your startup experience.
Business interview. This checks whether you are genuinely interested in the customer's workflow and business problems. In preparation you can first learn the basic insurance underwriting process and which pain points automation can solve, and in the interview ask lots of questions to show your curiosity.
Discussion
Loading comments…