BEGIN:VCALENDAR
VERSION:2.0
PRODID:IEEE vTools.Events//EN
CALSCALE:GREGORIAN
BEGIN:VTIMEZONE
TZID:America/New_York
BEGIN:DAYLIGHT
DTSTART:20260308T030000
TZOFFSETFROM:-0500
TZOFFSETTO:-0400
RRULE:FREQ=YEARLY;BYDAY=2SU;BYMONTH=3
TZNAME:EDT
END:DAYLIGHT
BEGIN:STANDARD
DTSTART:20261101T010000
TZOFFSETFROM:-0400
TZOFFSETTO:-0500
RRULE:FREQ=YEARLY;BYDAY=1SU;BYMONTH=11
TZNAME:EST
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
DTSTAMP:20260620T170551Z
UID:C4BFBBB2-558E-4DFC-8494-3500FCFBD344
DTSTART;TZID=America/New_York:20260725T100000
DTEND;TZID=America/New_York:20260725T130000
DESCRIPTION:At AWS scale\, a single capacity planning error misallocates hu
 ndreds of millions in CapEx. Spreadsheets don't survive at this velocity 
 — closed-loop AI systems do.\n\nThis session delivers a practitioner's u
 nfiltered view from inside hyperscale infrastructure planning at AWS. We e
 xamine how AI pipelines connect demand forecasting\, gap analysis\, build 
 planning\, and rack deployment into a single end-to-end system and what th
 at looks like in operational reality. We explore how feedback loops enable
  planning systems to detect model drift\, recalibrate on actuals\, and com
 pound forecast accuracy over time.\n\nBut intelligence without failure ana
 lysis is incomplete. We confront what happens when the forecast is wrong a
 t scale\, how failure cascades across procurement\, deployment\, and custo
 mer commitments\, and how human-in-the-loop governance and explainable AI 
 contain the damage and restore trust.\n\nAI that plans\, connects\, fails\
 , and learns.\n\nSpeaker(s): Nainsi Jain\, \n\nAgenda: \nArlington Central
  Library\n\n10 am - 11 am ----> Setup and Networking\n\n11 am -12 noon ---
 -> Speaker Presentation\n\n12 pm - 12:30 pm ---> Wrap up\n\n1015 N Quincy 
 St\, Arlington\, Airlington\, Virginia\, United States\, 22201\, Virtual: 
 https://events.vtools.ieee.org/m/560012
LOCATION:1015 N Quincy St\, Arlington\, Airlington\, Virginia\, United Stat
 es\, 22201\, Virtual: https://events.vtools.ieee.org/m/560012
ORGANIZER:krishna.kandi@ieee.org
SEQUENCE:38
SUMMARY:Closed-Loop Intelligence: How AI Plans\, Connects\, Fails\, and Lea
 rns Across the Hyperscale Infrastructure Lifecycle
URL;VALUE=URI:https://events.vtools.ieee.org/m/560012
X-ALT-DESC:Description: <br /><p class="MsoNormal">At AWS scale\, a single 
 capacity planning error misallocates hundreds of millions in CapEx. Spread
 sheets don't survive at this velocity &mdash\; closed-loop AI systems do.<
 /p>\n<p class="MsoNormal">This session delivers a practitioner's unfiltere
 d view from inside hyperscale infrastructure planning at AWS. We examine h
 ow AI pipelines connect demand forecasting\, gap analysis\, build planning
 \, and rack deployment into a single end-to-end system and what that looks
  like in operational reality. We explore how feedback loops enable plannin
 g systems to detect model drift\, recalibrate on actuals\, and compound fo
 recast accuracy over time.</p>\n<p class="MsoNormal">But intelligence with
 out failure analysis is incomplete. We confront what happens when the fore
 cast is wrong at scale\, how failure cascades across procurement\, deploym
 ent\, and customer commitments\, and how human-in-the-loop governance and 
 explainable AI contain the damage and restore trust.</p>\n<p class="MsoNor
 mal">&nbsp\;</p>\n<p class="MsoNormal">AI that plans\, connects\, fails\, 
 and learns.</p><br /><br />Agenda: <br /><p>Arlington Central Library</p>\
 n<p>10 am - 11 am&nbsp\; &nbsp\;----&gt\;&nbsp\; Setup and Networking</p>\
 n<p>11 am -12 noon ----&gt\; Speaker Presentation</p>\n<p>12 pm - 12:30 pm
  ---&gt\; Wrap up</p>
END:VEVENT
END:VCALENDAR