DoorDash

What does a 20 step warmup do to a 13B run?

What tradeoff does Llama 3's overtraining show?

What does perplexity measure and why is it an insufficient LLM quality signal?

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

Give the playbook when a model provider announces a deprecation deadline on your pinned model

At MoE inference, which resource constraints apply?

What is the primary advantage of chain-of-thought (CoT) judge prompts over direct score prompts?

Design an LLM judge prompt with anchor examples and explain how they reduce score variance

Match Langfuse score sources to the workflow they enable

Your team is paying $40k/month on LLM tokens: most calls share a long system prompt + 8k tokens of few-shot examples. How does prompt caching help and what's the expected savings?