gpu memory

The interviewer wants real numbers, not 'use a big GPU.' Weights are fixed, the KV cache grows with load, and the lever order decides everything. Here is the back-of-envelope and the serving stack.

The interviewer wants you to know why a model that fits on no single GPU still trains, and to do the optimizer-state memory math that motivates sharding. Here is the parallelism taxonomy and the 16-bytes-per-parameter calculation.

A favorite at hardware-aware shops. The signal is understanding that attention's cost is memory traffic, not FLOPs, and that FlashAttention is an exact, IO-aware reordering, not an approximation. Here is the answer that shows you think about the memory hierarchy.

OOM is the most common wall in deep learning, and 'buy a bigger GPU' is the weakest answer. The signal is naming the memory consumers, knowing which one dominates, and matching the right lever to it.

Data, tensor, and pipeline parallelism all leave one sequence's activations on one device, so 200k+ token training hits a wall none of them can fix. The fourth axis shards the sequence itself, and the interview lives in the communication math.