Roadmaps / Rust / Interview topics
Reference
Interview topics
Use these for Rust, systems and backend roles. Each follow-up is the question a good interviewer asks next. The Staff line is what lifts the answer.
Phase 1
| Question | Follow-up | A Staff answer adds |
|---|---|---|
| What does ownership give you? | What does it not prevent? | The bugs that remain (leaks, deadlocks, logic) and how you test for them |
When is unsafe justified? | How do you contain it? | A small module, a written invariant for each block and Miri in CI |
Rc or Arc? RefCell or Mutex? | What does each cost? | The cost measured, and a rule for the codebase |
| How do you design an error type? | Library or application? | Typed errors in libraries, context at the edge and no panics on input |
System design prompt (20 minutes): a key-value store for device data at 50,000 writes a second. Outline: requirements and numbers (3 min), storage layout (5), memory and unsafe (5), failure cases (4), testing (3).
Phase 2
| Question | Follow-up | A Staff answer adds |
|---|---|---|
How does async work in Rust? | What drives a future? | The runtime's role and what a blocking call does to it |
| What is cancellation safety? | How do you test it? | A drop-at-every-await test and a rule for write paths |
| How do you apply backpressure? | What happens with an unbounded queue? | Bounded channels, limits per connection and a shed policy |
Why is select! risky? | What do you do about it? | Which branches are cancellation safe, written next to the code |
System design prompt (20 minutes): a proxy that rate-limits 10,000 connections. Outline: limits and timeouts (4 min), async structure (5), shutdown (4), failure and abuse (4), observability (3).
Phase 3
| Question | Follow-up | A Staff answer adds |
|---|---|---|
| How do you test concurrent code? | What does a stress test miss? | loom for orderings and a stress run for scale, each with its limit |
What does fsync guarantee? | What can still lose data? | Directory syncs, device caches and a crash test that proves the claim |
| How do you find why a service is slow? | What if the profile is flat? | Counters, traces and a measured change, not a guess |
| Sharding or one lock? | How do you decide? | A measured crossover under your contention |
System design prompt (20 minutes): a write-ahead log that survives a power cut. Outline: record format (4 min), commit policy (5), recovery (5), testing (4), cost (2).
Phase 4
| Question | Follow-up | A Staff answer adds |
|---|---|---|
| How do you vet third-party crates? | What about build scripts? | A budget, a policy file, an advisory check and a review of build.rs |
| How do you expose Rust to Python or C? | Who owns the memory? | A written rule per pointer, a test on both sides and an ABI check |
| When should a team adopt Rust? | When should it not? | A decision memo with the costs measured on a real component |
| How do you keep up with a six-week release cycle? | What if a dependency lags? | A schedule, a pinned toolchain, an MSRV policy and a trial branch |
System design prompt (20 minutes): a plan to move one C component to Rust. Outline: why and what success means (4 min), interface and ownership (5), migration steps (4), risks and testing (4), a stop rule (3).