BEGIN:VCALENDAR VERSION:2.0 PRODID:IEEE vTools.Events//EN CALSCALE:GREGORIAN BEGIN:VTIMEZONE TZID:America/Denver BEGIN:DAYLIGHT DTSTART:20260308T030000 TZOFFSETFROM:-0700 TZOFFSETTO:-0600 RRULE:FREQ=YEARLY;BYDAY=2SU;BYMONTH=3 TZNAME:MDT END:DAYLIGHT BEGIN:STANDARD DTSTART:20261101T010000 TZOFFSETFROM:-0600 TZOFFSETTO:-0700 RRULE:FREQ=YEARLY;BYDAY=1SU;BYMONTH=11 TZNAME:MST END:STANDARD END:VTIMEZONE BEGIN:VEVENT DTSTAMP:20260716T232547Z UID:8437BA74-8DC2-4889-9127-C4531CB6B79C DTSTART;TZID=America/Denver:20260826T180000 DTEND;TZID=America/Denver:20260826T193000 DESCRIPTION:Presentation: Arrowheads and Circles: Building Self-Evaluating Loops in the Team and in the Product\n\nAbstract: As AI systems take on mo re real work\, the hard problem shifts from producing output to knowing wh ether to trust it. Large Language Models (LLMs) made generation cheap\; th ey did not make trust cheap. This talk is about building self-evaluating l oops around AI\, guided by one rule at two levels: no participant grades w ork it has a stake in. First in the team\, in the engineering process\, th rough review by a second model prompted to disprove the first\, contracts that check the boundaries between components\, and pipelines that allow on ly one source of truth for each piece of data. Then in the product\, insid e Orion\, a collaborative AI analyst platform whose claims ship with autom atic checks\, tested against live data and replaced rather than quietly ed ited when they go stale. Real production results included\, failures and a ll. Attendees leave with a one-afternoon audit they can run on their own s ystems: which feedback signals actually change what the system does\, and which are just decoration.\n\nSpeaker(s): Mr. Abijith Ramachandran\, \n\nV irtual: https://events.vtools.ieee.org/m/568296 LOCATION:Virtual: https://events.vtools.ieee.org/m/568296 ORGANIZER:GOWANSJ@IEEE.ORG SEQUENCE:78 SUMMARY:CIR: Arrowheads and Circles: Building Self-Evaluating Loops in the Team and in the Product URL;VALUE=URI:https://events.vtools.ieee.org/m/568296 X-ALT-DESC:Description: <br /><p class="MsoNormal"><span style="font-size: 11.5pt\; line-height: 115%\; font-family: 'AAAAAB+Georgia'\,sans-serif\; m so-bidi-font-family: AAAAAB+Georgia\;">Presentation: </span><strong><span style="font-size: 11.5pt\; line-height: 115%\;">Arrowheads and Circles: Bu ilding Self-Evaluating Loops in the Team and in the Product </span></stron g></p>\n<p class="MsoNormal"><strong><span style="font-size: 11.5pt\; line -height: 115%\;">&nbsp\;</span></strong></p>\n<p><span style="font-size: 1 1.5pt\; line-height: 115%\; font-family: 'AAAAAB+Georgia'\,sans-serif\; ms o-fareast-font-family: Calibri\; mso-fareast-theme-font: minor-latin\; mso -bidi-font-family: AAAAAB+Georgia\; mso-ansi-language: EN-US\; mso-fareast -language: EN-US\; mso-bidi-language: AR-SA\;">Abstract: As AI systems tak e on more real work\, the hard problem shifts from producing output to kno wing whether to trust it. Large Language Models (LLMs) made generation che ap\; they did not make trust cheap. This talk is about building self-evalu ating loops around AI\, guided by one rule at two levels: no participant g rades work it has a stake in. First in the team\, in the engineering proce ss\, through review by a second model prompted to disprove the first\, con tracts that check the boundaries between components\, and pipelines that a llow only one source of truth for each piece of data. Then in the product\ , inside Orion\, a collaborative AI analyst platform whose claims ship wit h automatic checks\, tested against live data and replaced rather than qui etly edited when they go stale. Real production results included\, failure s and all. Attendees leave with a one-afternoon audit they can run on thei r own systems: which feedback signals actually change what the system does \, and which are just decoration.</span></p> END:VEVENT END:VCALENDAR