BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//pretalx//pretalx.com//pyconde-pydata-2026//speaker//LDZFAV
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-pyconde-pydata-2026-R7TT3E@pretalx.com
DTSTART;TZID=CET:20260415T165500
DTEND;TZID=CET:20260415T172500
DESCRIPTION:In RAG-based systems\, the main challenge is often not tuning t
 he LLM itself\, but making documents available in a form that can be retri
 eved reliably. In enterprise settings\, the dominant input format is still
  PDF\, ranging from text-heavy reports to slide decks\, scanned documents\
 , and visually dense presentations. \n\nTraditional document processing pi
 pelines rely on OCR and layout analysis to extract text\, followed by chun
 king and embedding. While this works well for text-heavy documents\, much 
 of the original structure is often lost—especially for presentations\, m
 ulti-column layouts\, and visually driven content. Images\, charts\, and d
 iagrams typically require separate processing\, increasing pipeline comple
 xity and fragility.\n\nRecent multi-modal embedding models enable a differ
 ent approach: embedding entire PDF pages directly as images. This preserve
 s layout\, visual hierarchy\, and embedded graphics in a single representa
 tion and significantly simplifies document ingestion. \n\nThis talk compar
 es classical OCR-based document processing pipelines with multi-modal page
  embeddings\, drawing on benchmarks conducted on real-world enterprise doc
 uments across different models. It highlights where this approach performs
  well\, where its limitations lie\, and how to design practical\, cost-awa
 re retrieval systems in Python.
DTSTAMP:20260412T141741Z
LOCATION:Ferrum [2nd Floor]
SUMMARY:Simplifying RAG Document Pipelines with Multimodal Embeddings - Arn
 e Grobrügge
URL:https://pretalx.com/pyconde-pydata-2026/talk/R7TT3E/
END:VEVENT
END:VCALENDAR