Bypassing the GPU Tax: Smart Layer-7 Routing with Small Language Models (SLMs)
-
Sagar Utekar
CrowdStrike
SRE
How does Meta process trillions of background tasks daily without wasting expensive GPU capacity? Unpack the distributed queue architecture powering their unpredictable AI workloads.
Starting at 3:44 · Watch from the beginning
Checking access…
Playback and chapters load privately for Free videos.
Jobs that call for the skills explored in this talk.