BEGIN:VCALENDAR VERSION:2.0 PRODID:IEEE vTools.Events//EN CALSCALE:GREGORIAN BEGIN:VTIMEZONE TZID:America/Los_Angeles BEGIN:DAYLIGHT DTSTART:20260308T030000 TZOFFSETFROM:-0800 TZOFFSETTO:-0700 RRULE:FREQ=YEARLY;BYDAY=2SU;BYMONTH=3 TZNAME:PDT END:DAYLIGHT BEGIN:STANDARD DTSTART:20261101T010000 TZOFFSETFROM:-0700 TZOFFSETTO:-0800 RRULE:FREQ=YEARLY;BYDAY=1SU;BYMONTH=11 TZNAME:PST END:STANDARD END:VTIMEZONE BEGIN:VEVENT DTSTAMP:20261002T101751Z UID:E40D8EF9-6E19-4F77-BB16-D56FA4110985 DTSTART;TZID=America/Los_Angeles:20260924T173000 DTEND;TZID=America/Los_Angeles:20260924T193000 DESCRIPTION:[]\n\nThe rapid growth of AI is driving increasingly large and complex infrastructure deployments. While GPUs and accelerators receive mu ch of the attention\, bringing AI infrastructure into production involves a much broader challenge. Compute\, networking\, firmware\, hardware contr ollers\, security validation\, and provisioning services must all work tog ether reliably before an AI workload can run.\n\nDrawing from real-world e xperience designing and operating hyperscale cloud infrastructure\, this t alk explores what happens behind the scenes when AI infrastructure is prov isioned and prepared for production. It examines why provisioning at scale becomes a distributed systems problem\, where hardware heterogeneity\, de pendencies\, partial failures\, version drift\, and recovery can significa ntly affect reliability and operational efficiency.\n\nThe session will sh are practical engineering patterns for addressing these challenges\, inclu ding dependency-aware orchestration\, safe parallel execution\, idempotent workflows\, hardware and firmware validation\, observability\, failure is olation\, and resilient recovery. It will also explore how these patterns help infrastructure teams move from simply automating individual tasks tow ard building reliable lifecycle systems that can operate across large hete rogeneous fleets.\n\nRather than focusing on a specific cloud provider or technology stack\, the talk presents practical lessons and architectural p rinciples that engineers can apply when designing and operating reliable A I and cloud infrastructure at scale.\n\nArun Anbumani is a Principal Engin eer at Oracle Cloud Infrastructure (OCI)\, where he designs and operates l arge-scale cloud infrastructure supporting AI and enterprise workloads. Hi s work focuses on distributed systems\, infrastructure provisioning\, plat form reliability\, and automation at hyperscale. Arun is an IEEE Senior Me mber and also contributes to the broader engineering community through tec hnical writing and conference reviewing.\n\nVirtual: https://events.vtools .ieee.org/m/577229 LOCATION:Virtual: https://events.vtools.ieee.org/m/577229 ORGANIZER:amritesh557@ieee.org SEQUENCE:11 SUMMARY:Behind the Scenes of AI Infrastructure: Provisioning\, Reliability\ , and Scale URL;VALUE=URI:https://events.vtools.ieee.org/m/577229 X-ALT-DESC:Description: <br /><p>&nbsp\;</p>\n<p><img src="https://events.v tools.ieee.org/vtools_ui/media/display/f69dc8d8-ba9e-417a-969d-5027ea0cbc4 8" alt=""></p>\n<p>&nbsp\;</p>\n<p>The rapid growth of AI is driving incre asingly large and complex infrastructure deployments. While GPUs and accel erators receive much of the attention\, bringing AI infrastructure into pr oduction involves a much broader challenge. Compute\, networking\, firmwar e\, hardware controllers\, security validation\, and provisioning services must all work together reliably before an AI workload can run.</p>\n<p>Dr awing from real-world experience designing and operating hyperscale cloud infrastructure\, this talk explores what happens behind the scenes when AI infrastructure is provisioned and prepared for production. It examines wh y provisioning at scale becomes a distributed systems problem\, where hard ware heterogeneity\, dependencies\, partial failures\, version drift\, and recovery can significantly affect reliability and operational efficiency. </p>\n<p>The session will share practical engineering patterns for address ing these challenges\, including dependency-aware orchestration\, safe par allel execution\, idempotent workflows\, hardware and firmware validation\ , observability\, failure isolation\, and resilient recovery. It will also explore how these patterns help infrastructure teams move from simply aut omating individual tasks toward building reliable lifecycle systems that c an operate across large heterogeneous fleets.</p>\n<p>Rather than focusing on a specific cloud provider or technology stack\, the talk presents prac tical lessons and architectural principles that engineers can apply when d esigning and operating reliable AI and cloud infrastructure at scale.</p>\ n<p>&nbsp\;</p>\n<p><em>Arun Anbumani is a Principal Engineer at Oracle Cl oud Infrastructure (OCI)\, where he designs and operates large-scale cloud infrastructure supporting AI and enterprise workloads. His work focuses o n distributed systems\, infrastructure provisioning\, platform reliability \, and automation at hyperscale. Arun is an IEEE Senior Member and also co ntributes to the broader engineering community through technical writing a nd conference reviewing.</em></p> END:VEVENT END:VCALENDAR