DoorDash

Your nightly index rebuild takes 14 hours on CPU and the maintenance window is 2 hours. Pick the most direct fix.

Design an auto tuner that adjusts ef_search to hold recall@10 at 0.95 as data and traffic drift.

Calculate the per token KV cache footprint of a Llama-3-8B style config in fp16

Predict what a 20 step warmup does to a 13B run

Describe Guardrails AI's validator hub model

Decide between synchronous blocking and asynchronous logging for a new guardrail

Offline eval scores 0.9 but users report frequent wrong answers: why the gap, and what do you do?

When does an off the shelf embedding model fall short, and what is the fix?

Keeping an image heavy multimodal feature feeling fast despite a slow vision pipeline

Walk through the playbook when a model vendor announces a deprecation deadline on your pinned model