BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//pretalx//pretalx.devconf.info//devconf-us-2026//speaker//T7BG3Y
BEGIN:VTIMEZONE
TZID:EST
BEGIN:STANDARD
DTSTART:20001029T030000
RRULE:FREQ=YEARLY;BYDAY=-1SU;BYMONTH=10;UNTIL=20061029T070000Z
TZNAME:EST
TZOFFSETFROM:-0400
TZOFFSETTO:-0500
END:STANDARD
BEGIN:STANDARD
DTSTART:20071104T030000
RRULE:FREQ=YEARLY;BYDAY=1SU;BYMONTH=11
TZNAME:EST
TZOFFSETFROM:-0400
TZOFFSETTO:-0500
END:STANDARD
BEGIN:DAYLIGHT
DTSTART:20000402T030000
RRULE:FREQ=YEARLY;BYDAY=1SU;BYMONTH=4;UNTIL=20060402T080000Z
TZNAME:EDT
TZOFFSETFROM:-0500
TZOFFSETTO:-0400
END:DAYLIGHT
BEGIN:DAYLIGHT
DTSTART:20070311T030000
RRULE:FREQ=YEARLY;BYDAY=2SU;BYMONTH=3
TZNAME:EDT
TZOFFSETFROM:-0500
TZOFFSETTO:-0400
END:DAYLIGHT
END:VTIMEZONE
BEGIN:VEVENT
UID:pretalx-devconf-us-2026-X7VRBR@pretalx.devconf.info
DTSTART;TZID=EST:20260924T164000
DTEND;TZID=EST:20260924T171500
DESCRIPTION:Building a risk engine means making decisions from information 
 that is rarely complete\, clean or perfectly reliable.\n\nThe data can com
 e from APIs\, databases\, documents\, public sources\, historical events o
 r unstructured text. Some information may be missing\, contradictory\, out
 dated or even intentionally misleading.\n\nAdding AI to this process creat
 es new possibilities\, but also new engineering challenges. How much shoul
 d we trust the model? How do we validate its output? And more importantly\
 , how do we explain why the system reached a particular risk decision?\n\n
 In this talk\, I’ll share the engineering patterns behind building a pro
 duction-grade AI risk engine.\n\nI’ll cover:\n\n• How to design a pipe
 line that transforms heterogeneous and unstructured data into usable risk 
 signals\n\n• How to combine deterministic rules\, traditional data proce
 ssing and LLMs without making the LLM the source of truth\n\n• How to ha
 ndle missing\, conflicting and suspicious information\n\n• How to use co
 nfidence scoring and validation layers around AI generated signals\n\n• 
 How to aggregate multiple signals into an explainable risk decision\n\n•
  How to make every decision traceable and auditable\n\n• How to evaluate
  AI components when false positives and false negatives have real conseque
 nces\n\n• How to build observability around the whole pipeline so engine
 ers can understand what happened when a decision goes wrong
DTSTAMP:20260924T235747Z
LOCATION:Ladd Room (Capacity 170)
SUMMARY:From Messy Data to Explainable Decisions: Building a Production-Gra
 de AI Risk Engine - Anass Tissirallah
URL:https://pretalx.devconf.info/devconf-us-2026/talk/X7VRBR/
END:VEVENT
END:VCALENDAR