BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//pretalx//pretalx.com//pyconde-pydata-2026//talk//F79RG9
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-F79RG9@pretalx.com
DTSTART;TZID=CET:20260415T150000
DTEND;TZID=CET:20260415T154500
DESCRIPTION:This talk will detail how we used Rust to solve a number of res
 ource utilization inefficiencies while scaling data pre-processing to a pe
 tabyte scale and enable next-generation model training at DeepL. Besides o
 ther factors\, this was done by developing an internal library for interac
 ting with Parquet files in a memory efficient nature.\n\nTopics include:\n
 • Convincing you to love Rust for its memory safety\n• Comparing C++ a
 nd Rust ecosystems for Python library development\n• Diving into Python-
 Rust interoperability\n• Convincing you to love Rust for its user-friend
 ly (yes\, actually!) language features\n• Providing a high-level overvie
 w of the continuously growing impact that Rust is having on the Arrow and 
 data engineering ecosystem
DTSTAMP:20260615T085525Z
LOCATION:Platinum [2nd Floor]
SUMMARY:Scaling Data Processing for Training Workloads at DeepL Research wi
 th Rust - Jonas Dedden\, Johanna Goergen
URL:https://pretalx.com/pyconde-pydata-2026/talk/F79RG9/
END:VEVENT
END:VCALENDAR