Interview Prep — Head of Software Engineering [GPU]

This is the interview surface for the role, drawn from the 12 phases. The role is senior and broad, so interviews span deep technical, systems design, and leadership — often in the same loop.

The five question banks

FileCoversPhases
01 — GPU & Systems Deep-Divearchitecture, CUDA, compilers, drivers, runtime01–05
02 — Serving, Scheduling & DistributedKV-cache, batching, GPU sharing, collectives06–08
03 — Platform, Portability & SecurityHAL/ABI, observability, licensing, sovereign09–11
04 — System Designthe big design walkthroughsall
05 — Leadership & Behavioralstrategy, org, OEM, communication, STAR12

How to use this

  1. Self-test first: answer cold, then check against the model answer + the WARMUP it derives from. Gaps point you back to a phase.
  2. The bar is "derive, don't recite": every technical answer should reason from a principle (roofline, the 2V all-reduce bound, sign-vs-encrypt), not a memorized fact. The WARMUPs' Interview Q&A sections are the depth target.
  3. Leadership answers need a decision + its consequence + the honest limit — the Phase 12 communication discipline.

The 20 questions most likely to decide the loop

  1. Derive H100's ridge point and what it means for LLM decode. (P01)
  2. Walk me through what happens at the OS level on cudaLaunchKernel. (P02/P04)
  3. What does nvcc do? PTX vs SASS, and the cold-start incident. (P03)
  4. Why does PyTorch ship its own allocator? Reserved vs allocated. (P05)
  5. Compare MPS, MIG, time-slicing; deliver a "0.5 GPU" tier honestly. (P06)
  6. Explain PagedAttention — and why it's really an allocator. (P07/P05)
  7. Static vs continuous batching; what is goodput and why does it beat throughput? (P07)
  8. Derive ring all-reduce's cost; why is it bandwidth-optimal and N-independent? (P08)
  9. Why must tensor parallelism stay on NVLink? With numbers. (P08)
  10. Design a hardware-abstraction layer; walk the API/ABI boundary. (P09)
  11. How do you avoid a lowest-common-denominator abstraction? (P09)
  12. How do you version a plugin ABI so a year-old vendor backend still loads? (P09)
  13. What GPU metrics do you alert on, and why is "GPU utilization" misleading? (P10)
  14. Design a safe release process for a GPU serving platform. (P10)
  15. Design a license system for air-gapped deployment; sign vs encrypt? (P11)
  16. Anti-tamper self-checks — what do they buy, how are they defeated? (P11)
  17. Build vs adopt: your own compiler layer or Triton/MLIR? Defend it. (P03/P09)
  18. How would you execute "decouple software from hardware"? (P09/P12)
  19. How do you structure and hire a GPU platform team? (P12)
  20. Sell me the sovereign-AI value proposition. (P12)

Each maps to a WARMUP's Interview Q&A with a full principal-level answer.

Format expectations for this role

  • Technical screens: whiteboard a kernel/scheduler/allocator; reason about a roofline; debug a "slow training" or "slow serving" scenario.
  • System design: the system-design walkthroughs (GPU cloud, serving platform, HAL, licensing-at-scale).
  • Leadership: strategy memos, org design, OEM/partner scenarios, and STAR behavioral stories — see 05.
  • Hands-on credibility check: expect to be asked to actually reason about code you'd review — the labs are your evidence you can.