Advanced Deep Learning Interview Questions, issue 2, Mar 23, 2026

The Memory Fragmentation Trap

Senior ML Engineer interview at Meta, and the interviewer asks:

A junior dev hands you a 500-line PyTorch Out-of-Memory (OOM) stack trace and asks for help. What is your exact debugging workflow before you even think about telling them to ‘just lower the batch size’?

Don’t say: I’d look at the bottom of the trace to find the failing layer, check the tensor shapes, and maybe suggest turning on gradient checkpointing.

OOMs aren’t always about capacity - they’re often allocator fragmentation failures that only surface under peak allocation pressure.

The full answer, with the mechanism and the arithmetic, is for paid subscribers on Substack.

Read it on Substack

Get the next one

Free on Substack. Unsubscribe in one click.