BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//pretalx//pretalx.com//pydata-amsterdam2026//speaker//BPMEGF
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-pydata-amsterdam2026-SPZLFK@pretalx.com
DTSTART;TZID=CET:20260910T142500
DTEND;TZID=CET:20260910T145500
DESCRIPTION:Current LLM workflows work surprisingly well until context beco
 mes the bottleneck. Tasks like multi-file code changes\, long-log debuggin
 g\, or stateful tool use that many teams attempt to solve with prompt engi
 neering or agents often succeed at small scale and then become brittle as 
 the amount of context grows.\n\nThis talk introduces Recursive Language Mo
 dels (RLMs)\, a different approach in which context is treated as data tha
 t can be programmatically explored\, decomposed\, and revisited - rather t
 han input that must be consumed in a single prompt. This shifts long-conte
 xt LLM systems from brittle prompt orchestration to programs that explicit
 ly explore\, track\, and update context over time.\n\nUsing DSPy as a conc
 rete Python implementation\, I will demonstrate how this approach turns a 
 failing long-context task into a tractable one\, and what this shift means
  for designing LLM systems in practice.\n\nAttendees will leave with a cle
 ar mental model for when standard prompting breaks\, what RLMs change tech
 nically\, and how to prototype this pattern in practice.
DTSTAMP:20260710T154856Z
LOCATION:Room 2 (350)
SUMMARY:When Context Breaks: Recursive Language Models with DSPy - Niels va
 n Galen Last
URL:https://pretalx.com/pydata-amsterdam2026/talk/SPZLFK/
END:VEVENT
END:VCALENDAR