BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//pretalx//pretalx.devconf.info//devconf-cz-2026//talk//MCWGNQ
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-devconf-cz-2026-MCWGNQ@pretalx.devconf.info
DTSTART;TZID=CET:20260618T144500
DTEND;TZID=CET:20260618T152000
DESCRIPTION:There is a massive gap between a flashy prototype and an AI ser
 vice you can trust. While implementing an agentic service to automate RPM 
 packaging\, we learned that while AI handles complex tasks\, your code mus
 t provide reliability. The real challenge was building a "safety harness" 
 to keep a non-deterministic model from breaking deterministic systems.\nTh
 is talk shares the practical engineering required to make agents productio
 n-ready:\n- Sandboxing: Why you should never let an agent run free and how
  to use isolated environments to ensure AI commands can't damage your syst
 em.\n- Validation loops: Automated checks that treat every AI suggestion a
 s an untrusted draft that must be verified before execution.\n- Observabil
 ity: Move beyond the "black box" by tracking an agent’s "train of though
 t" so you can debug AI failures like any other software bug.\n\nIf you wan
 t to build AI tools as stable\, observable\, and secure as traditional cod
 e\, this session provides the blueprint we used to get there.
DTSTAMP:20260430T131652Z
LOCATION:A112 (capacity 64)
SUMMARY:How AI helped us ship updates in a Linux distro - Tomas Tomecek\, L
 aura Barcziova
URL:https://pretalx.devconf.info/devconf-cz-2026/talk/MCWGNQ/
END:VEVENT
END:VCALENDAR