BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//pretalx//program.berlinbuzzwords.de//bbuzz26//speaker//8TWA3E
BEGIN:VTIMEZONE
TZID:CET
BEGIN:STANDARD
DTSTART:20001029T040000
RRULE:FREQ=YEARLY;BYDAY=-1SU;BYMONTH=10
TZNAME:CET
TZOFFSETFROM:+0200
TZOFFSETTO:+0100
END:STANDARD
BEGIN:DAYLIGHT
DTSTART:20000326T030000
RRULE:FREQ=YEARLY;BYDAY=-1SU;BYMONTH=3
TZNAME:CEST
TZOFFSETFROM:+0100
TZOFFSETTO:+0200
END:DAYLIGHT
END:VTIMEZONE
BEGIN:VEVENT
UID:pretalx-bbuzz26-GPKCWA@program.berlinbuzzwords.de
DTSTART;TZID=CET:20260609T140000
DTEND;TZID=CET:20260609T144000
DESCRIPTION:Competitive search now needs dense embeddings\, sparse vectors\
 , ColBERT\, and cross-encoder reranking. Most teams run four separate cont
 ainers. This talk shows how to serve all four from one process\, walks thr
 ough building a hybrid retrieval pipeline with real benchmark data\, and c
 overs where each retrieval mode wins and where it wastes compute.
DTSTAMP:20260525T114315Z
LOCATION:Maschinenhaus
SUMMARY:One GPU\, Four Retrieval Modes: Multi-Model Search Serving - Filip 
 Makraduli
URL:https://program.berlinbuzzwords.de/bbuzz26/talk/GPKCWA/
END:VEVENT
END:VCALENDAR