BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//CERN//INDICO//EN
BEGIN:VEVENT
SUMMARY:STEAM Academy Seminar: The Secrets to Training World-Class LLMs
DTSTART:20260820T140000Z
DTEND:20260820T153000Z
DTSTAMP:20260810T085800Z
UID:indico-event-1696055@indico.cern.ch
CONTACT:steam.academy@cern.ch\;ep-seminars.colloquia@cern.ch
DESCRIPTION:Speakers: Lewis Tunstall (Hugging Face)\n\nAbstract: What does
  it actually take to train a strong language model? Published research mak
 es it look clean: strategic architecture choices\, carefully curated data\
 , enough compute\, and a tidy set of ablations where every decision seems 
 obvious in hindsight. But those reports are written with a fair bit of ros
 y retrospection. They don't capture the 2 a.m. debugging sessions\, the lo
 ss spikes that appear from nowhere\, or the subtle tensor-parallelism bug 
 that quietly sabotages a run for days. The reality is messier and more ite
 rative\, and most of the decisions that actually mattered never make it in
 to the final paper. In this lecture I'll walk through the nuts and bolts o
 f training an LLM end to end\, from pretraining over trillions of tokens t
 o post-training with reinforcement learning. Speaker's Bio: Lewis is a M
 achine Learning Engineer at Hugging Face where he works on applying Transf
 ormers to automate business processes and solve MLOps challenges. Lewis ha
 s built ML applications for startups and enterprises in the domains of NLP
 \, topological data analysis\, and time series. In a previous life\, Lewis
  was a theoretical physicist and contributer to open-source projects.\nThi
 s seminar is part of the CERN STEAM Academy Seminar Series. \nNetworking 
 cocktail will follow the seminar. \n\nWith the support of CERN's Next Ge
 neration Triggers Project.\n\n \n\n\nhttps://indico.cern.ch/event/1696055
 /
LOCATION:500/1-001 - Main Auditorium (CERN)
URL:https://indico.cern.ch/event/1696055/
END:VEVENT
END:VCALENDAR