Loading
Is Speculative Decoding Worth It? Profiling vLLM on NVIDIA Blackwell — Akamai
Akamai's Sheilah Kirui benchmarks speculative decoding with vLLM on a Blackwell GPU and shows when a draft model pays off (1.6x on structured output) and when it doesn't.
More in Infra
1:51:37InfraAdvanced Workshop: Mastering AI Observability — Doug Guthrie, Braintrust
BraintrustobservabilityLLMOps
1:20:35InfraWhat If We Stopped Using GPUs? | YC Paper Club
YCGPUsinference
59:59InfraWhat Is an Inference Engine, Anyway? — Charles Frye, Modal
Modalinferenceserving
