Thursday, 9 October, 2025 - 09:25–09:45
John Lunney, Google
Accelerators for serving LLMs are a very scarce resource, both globally and inside your organization. You must show you're making good use of the resources, otherwise someone else will.
John Lunney is a Senior Staff Reliability Engineer at Google. He is the technical lead for Workspace AI SRE, running a platform for LLM-powered features. He holds a degree in Computational Linguistics from Trinity College in Dublin, Ireland. Before Google, he worked on several lexicography projects for the Irish language. [email protected]
BibTeX
@conference {311896,
author = {John Lunney},
title = {Maximizing Utilization for {LLM} Accelerators},
year = {2025},
address = {Dublin},
publisher = {USENIX Association},
month = oct
}
author = {John Lunney},
title = {Maximizing Utilization for {LLM} Accelerators},
year = {2025},
address = {Dublin},
publisher = {USENIX Association},
month = oct
}
