Interview process
- Overall it was a technical interview lasting about 1 hour
- The interviewer was pretty friendly, but the questions covered a wide range — felt like they were testing general SRE skills plus coding fundamentals
Interview content
- Kubernetes
- Asked some basic K8s principles, like the difference between Pod and Deployment, and what a Service does.
- Also asked how to debug a Pod that fails to start.
- Python trivia
- Asked common Python questions, like mutable vs immutable objects, shallow copy vs deep copy, what the GIL does, etc.
- SLO (Service Level Objective)
- Had me explain the difference between SLO, SLA, and SLI.
- Asked what to do if the actual error rate exceeds the SLO.
- Distributed System Troubleshooting
- Gave a scenario where request latency in a distributed system was very high, and asked which areas to investigate.
- Things like network, load balancer, DB, cache, the service itself, etc.
- NoSQL vs SQL
- Asked about the difference between NoSQL and SQL, and when to choose NoSQL.
- I answered with schema flexibility, horizontal scalability, and suitability for high-concurrency/massive-data scenarios.
- LeetCode Coding
- Tested LeetCode #94: Binary Tree Inorder Traversal.
- I wrote both the iterative and recursive solutions.
Overall impressions
- TikTok's SRE interview leans more toward a Dev + Ops combination — you need coding ability, but you also need to understand distributed systems and ops principles.
- I feel like if you're interviewing for SWE instead, it's probably more algorithm- and system-design-heavy.
- My advice for anyone preparing: review K8s basics, Python trivia, core SRE concepts (SLO/SLI/SLA), distributed troubleshooting approaches, and grind some basic LeetCode binary tree/linked list problems.
Discussion
Loading comments…