BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//pretalx//talks.devopsdays.org//devopsdays-portland-2026//speaker
 //XGZRGV
BEGIN:VTIMEZONE
TZID:US/Pacific
BEGIN:DAYLIGHT
DTSTART:20250909T000000
TZNAME:PDT
TZOFFSETFROM:-0700
TZOFFSETTO:-0700
END:DAYLIGHT
BEGIN:STANDARD
DTSTART:20251102T020000
RDATE:20261101T020000
TZNAME:PST
TZOFFSETFROM:-0700
TZOFFSETTO:-0800
END:STANDARD
BEGIN:DAYLIGHT
DTSTART:20260308T030000
RDATE:20270314T030000
TZNAME:PDT
TZOFFSETFROM:-0800
TZOFFSETTO:-0700
END:DAYLIGHT
END:VTIMEZONE
BEGIN:VEVENT
SUMMARY:So You Want to DevOps an LLM: A Field Guide from Confusion to Clar
 ity - Maarika Teose
DTSTART;TZID=US/Pacific:20260909T104000
DTEND;TZID=US/Pacific:20260909T111000
DTSTAMP:20260824T030026Z
UID:pretalx-devopsdays-portland-2026-HJUQWB@talks.devopsdays.org
DESCRIPTION:What does it actually take to run LLMs in production and why i
 s it so hard to find clear answers?\nThis talk is a practical walkthrough 
 of the core MLOps concepts behind modern LLM systems\, built from scratch 
 out of necessity. Faced with a scattered ecosystem and limited end-to-end 
 guidance\, I set out to piece together how everything fits: from GPUs and 
 custom silicon\, to ML pipelines and fine-tuning\, to the realities of ser
 ving models at scale.\nIn 30 minutes\, I will try to lay out the fundament
 als\, such as hardware tradeoffs\, model optimization techniques like quan
 tization\, and the mechanics of high-performance inference such as KV cach
 ing\, batching\, and distributed serving. \nThis isn’t a polished framew
 ork or step-by-step guide\, but it is a cohesive map of the basics. If you
 ’ve struggled to understand how to “DevOps” LLMs\, this talk will he
 lp you bridge the gaps and build your own mental model.
LOCATION:Ballroom
URL:https://talks.devopsdays.org/devopsdays-portland-2026/talk/HJUQWB/
END:VEVENT
END:VCALENDAR