I went through Accenture's loop targeting their Technology Senior Manager track, which is roughly the senior/principal IC equivalent if you're coming from a pure engineering background. Here's what the system design round actually looked like.
The prompt: design a notification service for a large enterprise client. Millions of events per day, multiple channels (email, SMS, push, webhook). They explicitly told me to treat this like I was designing it for a real client engagement.
That framing matters. This isn't a FAANG-style "design Twitter" where clever distributed systems knowledge wins the round. At Accenture, the interviewer cared more about: How I would communicate trade-offs to a non-technical client stakeholder How the architecture fits into an existing enterprise ecosystem (SAP, Salesforce, legacy ESBs came up) Operational concerns: monitoring, SLAs, escalation paths Cost. They actually asked about cloud cost trade-offs between SQS/SNS on AWS vs. a self-hosted Kafka setup. Real dollars.
The depth on distributed systems was lighter than what I'd expect at a Google or Meta loop. They weren't looking for me to optimize tail latency at the 99.9th percentile. They wanted to know I could design something maintainable and explainable.
What tripped me up was forgetting to account for client-side constraints. They asked: "What if the client's security team won't allow outbound webhooks?" I had to pivot the design mid-round and that felt awkward. Point being: think about enterprise procurement constraints, not just technical ones.
Time: 45 minutes. Interviewer was a principal engineer with a lot of delivery background, not a pure research type.
For prep: I'd prioritize system design communication over raw architecture depth. Know how to draw a clean diagram, explain it in plain English, and handle scope pushback.