As artificial intelligence moves out of the lab and into production, the technology industry is figuring out what it means to run AI inference workloads at scale. Power concerns, multi-agent systems, and novel security implications all complicate the situation — meaning it requires much more strategy than simply having ample budget for GPUs.
In this video, Anat Heilper, director of AI architecture at VAST Data, explains, among other things, how production AI has evolved, the role of KVcache for workload efficiency, and where to begin thinking about capacity planning for large-scale agentic deployments.
If you want to hear more from Anat, watch the full video here, or check out her recent blog posts diving into topics such as KVcache tiering and agentic infrastructure:
