Went through the KP system design loop last month for a staff-level role. I've done enough of these at this point (FAANG, multiple mid-size healthtech, now this) that I can calibrate where KP sits.
Short answer: the bar is real but it's not an L6 Google loop. The depth they want is roughly senior IC at a large enterprise. They care about different things than pure tech companies.
What they actually asked (paraphrasing): Design a system that delivers real-time medication reminders to patients across multiple channels (SMS, push, in-app). Patient population is millions. What happens if a channel fails?
That's the flavor. Healthcare-domain-flavored distributed systems. You need to talk about: Reliability and fallback strategies (they care a lot about what happens when things break) PHI handling and why it matters architecturally (not just 'encrypt everything', but where encryption lives, who can decrypt, logging) Scalability across a huge member base (KP covers 12+ million members) Integration with legacy systems, because they have a lot of them
They were not impressed by buzzword-dropping. When I mentioned Kafka, the interviewer immediately asked follow-up questions about consumer lag, ordering guarantees, and how I'd handle poison pill messages. Bring receipts.
One thing that caught me: they asked about regulatory constraints as a first-class design consideration. Not an afterthought. Something like 'what constraints does HIPAA put on your logging design?' You don't need to be a lawyer, but you should understand that PHI can't just go into your standard CloudWatch logs without masking.
Time allocation in the interview: about 10 minutes scoping/clarifying, 25 minutes building the design, 10 minutes on deep-dives of their choosing. Pretty standard.
The interviewer was a senior principal, asked good questions, didn't try to trick me. Felt like a real conversation.