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:20260612T084551Z
UID:F37FC9C6-17B8-4EF7-8A02-C79D2E23F8EE
DTSTART;TZID=America/Los_Angeles:20260715T180000
DTEND;TZID=America/Los_Angeles:20260715T190000
DESCRIPTION:As artificial intelligence transitions from experimental deploy
 ments to mission-critical production infrastructure\, organizations face a
  fundamental shift from model-centric optimization to system-centric engin
 eering. While advances in model architectures and accelerator technologies
  have driven recent AI breakthroughs\, long-term performance\, reliability
 \, and sustainability increasingly depend on the interaction between compu
 te\, memory\, networking\, software runtimes\, operations\, and governance
 . This talk presents an industry roadmap for building scalable AI systems 
 that move beyond isolated model optimization toward adaptive\, software-de
 fined AI platforms. The roadmap explores five interconnected layers—comp
 ute\, memory and data\, interconnect\, runtime and operating systems\, and
  operations and governance—and demonstrates how these layers collectivel
 y influence throughput\, latency\, cost\, energy efficiency\, reliability\
 , and compliance. The discussion introduces workload-aware architectures f
 or inference\, retrieval-augmented generation (RAG)\, agentic workflows\, 
 multimodal applications\, and edge AI\, highlighting the growing importanc
 e of memory hierarchies\, topology-aware scheduling\, adaptive control loo
 ps\, and cluster-scale orchestration. A practical AI systems maturity mode
 l is proposed to help organizations assess current capabilities and priori
 tize investments\, progressing from ad hoc experimentation to autonomous\,
  policy-governed AI fabrics. The presentation concludes with a pragmatic e
 xecution framework and industry best practices for achieving predictable s
 ervice levels\, operational resilience\, and sustainable AI economics. The
  central thesis is that future AI leadership will be determined not by mod
 el performance alone\, but by the ability to design\, operate\, and govern
  AI as an integrated systems platform\n\nCo-sponsored by: Vishnu S. Pendya
 la\, San Jose State University\n\nSpeaker(s): Sujit Reddy Thumma\n\nVirtua
 l: https://events.vtools.ieee.org/m/563401
LOCATION:Virtual: https://events.vtools.ieee.org/m/563401
ORGANIZER:rahul.110392@gmail.com
SEQUENCE:0
SUMMARY:Beyond the Model: The Industry Playbook for Scalable AI Systems
URL;VALUE=URI:https://events.vtools.ieee.org/m/563401
X-ALT-DESC:Description: <br /><p class="p1">As artificial intelligence tran
 sitions from experimental deployments to mission-critical production infra
 structure\, organizations face a fundamental shift from model-centric opti
 mization to system-centric engineering. While advances in model architectu
 res and accelerator technologies have driven recent AI breakthroughs\, lon
 g-term performance\, reliability\, and sustainability increasingly depend 
 on the interaction between compute\, memory\, networking\, software runtim
 es\, operations\, and governance. This talk presents an industry roadmap f
 or building scalable AI systems that move beyond isolated model optimizati
 on toward adaptive\, software-defined AI platforms. The roadmap explores f
 ive interconnected layers&mdash\;compute\, memory and data\, interconnect\
 , runtime and operating systems\, and operations and governance&mdash\;and
  demonstrates how these layers collectively influence throughput\, latency
 \, cost\, energy efficiency\, reliability\, and compliance. The discussion
  introduces workload-aware architectures for inference\, retrieval-augment
 ed generation (RAG)\, agentic workflows\, multimodal applications\, and ed
 ge AI\, highlighting the growing importance of memory hierarchies\, topolo
 gy-aware scheduling\, adaptive control loops\, and cluster-scale orchestra
 tion. A practical AI systems maturity model is proposed to help organizati
 ons assess current capabilities and prioritize investments\, progressing f
 rom ad hoc experimentation to autonomous\, policy-governed AI fabrics. The
  presentation concludes with a pragmatic execution framework and industry 
 best practices for achieving predictable service levels\, operational resi
 lience\, and sustainable AI economics. The central thesis is that future A
 I leadership will be determined not by model performance alone\, but by th
 e ability to design\, operate\, and govern AI as an integrated systems pla
 tform</p>
END:VEVENT
END:VCALENDAR