Company vault
GPU and AI compute — C++/CUDA, performance, and systems in Santa Clara and Austin.
Questions
105
Reported in NVIDIA interview loops. Heap and selection strategy question with multiple valid approaches.
Reported in NVIDIA interview loops. Advanced string window question seen in strong coding interviews.
Reported in NVIDIA interview loops. Graph traversal question using DFS or BFS flood-fill.
Reported in NVIDIA interview loops. Topological sort and directed cycle detection interview staple.
Reported in NVIDIA interview loops. Dynamic programming optimization with unbounded choices.
Reported in NVIDIA interview loops. Simple but important recurrence-based DP interview problem.
Reported in NVIDIA interview loops. Data structure trade-off question for autocomplete and dictionary lookups.
Reported in NVIDIA interview loops. Hard interview classic requiring binary search partition logic.
Reported in NVIDIA interview loops. Disjoint set union structure used in graph connectivity interviews.
Reported in NVIDIA interview loops. Pattern recognition question for linear-time array scans.
Reported in NVIDIA interview loops. System design problem testing reliability, fan-out, and channel abstraction.
Reported in NVIDIA interview loops. Design question around large files, durability, and secure uploads.
Reported in NVIDIA interview loops. System design interview around indexing, ranking, and freshness.
Reported in NVIDIA interview loops. Critical design topic focused on correctness under retries and failures.
Reported in NVIDIA interview loops. Geo-spatial system design covering matching latency and surge logic.
Reported in NVIDIA interview loops. Observability architecture interview question for platform roles.
Reported in NVIDIA interview loops. System design exercise emphasizing control plane and low-latency reads.
Reported in NVIDIA interview loops. Architecture discussion around tenant security and scaling models.
Reported in NVIDIA interview loops. System design topic around edge caching, invalidation, and latency.
Reported in NVIDIA interview loops. Event delivery design question focusing on retries and idempotency.