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:20261017T195839Z UID:F642CB7D-729B-4AA8-A864-730717FA3448 DTSTART;TZID=America/New_York:20261024T083000 DTEND;TZID=America/New_York:20261024T183000 DESCRIPTION:FREE REGISTRATION OPEN - 23rd OCT 2026\n\nIEEE Workshop on AI-N ative Networking\, Resilience and Forensics\nDETAILED PRELIMINARY PROGRAM\ nSaturday\, October 24\, 2026 | Carlson Auditorium\, Rochester Institute o f Technology\n8:30 AM–6:05 PM • Technical talks\, panels and student p oster awards\n\nOpening & Session 1 | AI Networking and Infrastructure\n\n 8:30–8:45 AM\nWelcome and Introduction — Nirmala Shenoy (Rochester Ins titute of Technology)\n\n8:50–9:20\nLoomAI and Autonomous Experiments on the FABRIC Testbed\nPaul Ruth | RENCI\, University of North Carolina at C hapel Hill\nUsing AI agents to provision\, operate\, analyze and reproduce distributed FABRIC experiments\, and using FABRIC to study agents across real networks.\n\n9:20–9:40\nPowering the Next-Gen Campus: The Impact of AI on Infrastructure and Networking Resources in GCCIS\nCharles Gruener | Director of Technical Operations\, GCCIS\, Rochester Institute of Technol ogy\nCampus computing and networking capacity\, orchestration and resource -management implications of emerging AI workloads.\n\n9:40–10:00\nSoftwa re and Network Challenges of Making GPUs Work Together at Scale\nDeep Shah | Socratic AI (formerly Meta and VMware)\nHow distributed AI training req uires GPU coordination\, network communication and mechanisms for performa nce and failure recovery.\n\n10:00–10:20\nMemory Is the Bottleneck: Pool ing GPUs for One Big Job vs. Slicing One GPU for Many Robots\nFrancis Tyso n Thomas | Oishii Farm\; formerly Kroll Government Solutions\nCompares poo ling GPUs across nodes for large jobs with sharing a GPU for real-time rob otic inference\, including latency and observability tradeoffs.\n\n10:20 –10:35\nMORNING COFFEE BREAK\n\nSession 2 | AI Optimization and Resilien ce\n\n10:40–11:00\nAdvanced Latent Representations and Computational Opt imization\nBaris Kazar | Infrastructure architect and spatial data scienti st\; formerly Oracle America\nOptimization of embeddings\, fused Triton/CU DA kernels\, vector quantization and approximate nearest-neighbor search f or AI systems.\n\n11:00–11:20\nKeeping Distributed AI Inference Reliable Under Overload: Deadline-Aware Backpressure and Adaptive Load Shedding\nJ ayapragash Dakshnamurthy | HealthEquity\nDA-BALS uses deadline and congest ion signals for admission and load shedding in distributed cloud–edge AI inference.\n\n11:20–11:40\nWhen the Retry Is the Outage: Metastable Fai lures in Agentic and Distributed AI Systems\nAbdul Rasheed Shaik | Senior engineering leader\, Fortune 100 firm\nSimulation-based analysis of retry amplification\, persistent overload and controls such as cancellation\, re try budgets and admission deadlines. Co-author: Khadarvali Shaik (Fortune 50 healthcare organization)\; Abdul presents.\n\n11:45–12:25 PM\nTECHNIC AL PANEL 1 | Networking for the AI Era?\nVidya Wadkar — Verizon\nSwapna Chimanchodkar — Independent Researcher\nNirmala Shenoy — Rochester Ins titute of Technology\nGuru Sharan B S — Founder & CEO\, Infi Tech World\ n\n12:25–1:15\nLUNCH BREAK & STUDENT POSTER AWARDS\n\nSession 3 | AI-Nat ive Infrastructure and Operations\n\n1:20–1:40 PM\nEmbedded System Devel opment in Industry: Before and After AI\nDr. Varma Kamadi | HCL America (H CL Technologies)\nPractical changes AI brings to embedded development\, fr om interpreting datasheets and writing drivers to debugging and testing fi rmware.\n\n1:40–2:00\nBolt-On Intelligence: Incrementally Engineering an AI Layer for Existing Network Infrastructure\nGabriel W. Nellis | Morgan Properties\; student\, Rochester Institute of Technology\nA staged approac h to network intelligence using management interfaces\, shared data models \, deterministic analysis and carefully governed LLM integration.\n\n2:00 –2:20\nAgentic AI for Predictive Failure Detection and Autonomous Recove ry in Cloud-Native Infrastructure\nRambabu Tangirala | Senior data enginee r and independent researcher\nCombining observability\, anomaly detection\ , retrieval-augmented generation and AI agents for root-cause analysis and human-approved recovery.\n\n2:20–2:40\nScaling Distributed Corporate In frastructures via AI-Native Architectures: An Empirical Framework for Adap tive Resource Allocation and Deterministic Compliance\nGuru Sharan B S | F ounder & CEO\, Infi Tech World\nEnterprise AI architectures using localize d vector sharding\, RAG\, multi-agent orchestration and policy-aware hybri d-cloud resource allocation.\n\n2:45–3:20\nTECHNICAL PANEL 2 | Trustwort hy Distributed AI Infrastructure: Security\, Observability and Forensics\n Prasanna Kumar — M&T Bank\;\nChitiz Tayal — Senior Director\, Axtria\n Venkata Phani Patelkhana — Independent researcher\nBenjamin Brümm — D alhousie University / Open Compute Project\n\n3:20–3:35\nAFTERNOON COFFE E BREAK\n\nSession 4 | AI Governance\, Security and Forensics\n\n3:40–4: 00 PM\nEvery Action Should Be Accounted For: Behavioral Forensics for AI A gents in Data Center Operations\nRea Ara | New York University (NYU Tandon )\nEnumerates agent execution paths and tests deterministic rules against security risks in data-center facilities operations.\n\n4:00–4:20\nGover ning Unreliable AI Services in Regulated Workflows\nAnurag Nedunuri | Inde pendent researcher\; lead solutions architect\nAuditable asynchronous orch estration of AI requests\, callbacks\, exceptions and human overrides in r egulated insurance workflows.\n\n4:20–4:40\nBeyond Suspicion Scores: Com peting-Explanation Inference for Fraud-Shaped Legitimate Activity\nAnkur G arg | Director of AI and Emerging Tech\, Rockland Trust Bank\nExplaining c ompeting fraud and legitimate-activity hypotheses to improve transparency and auditability in financial investigations.\n\n4:40–5:00\nEnterprise C ontext Layer: The Missing Foundation for Agentic AI\nChitiz Tayal | Senior Director\, Axtria\nWhy governed meaning\, permissions and enterprise cont ext must connect data platforms with agents to prevent context-driven erro rs.\n\n5:00–5:20\nFrom Request to Action: Inside the Journey of an AI Ag ent Decision\nVenkata Phani Patelkhana | Distinguished Engineer and indepe ndent researcher\nFollows user intent through agent identity\, delegated a uthority\, context\, policy checks\, tool execution and audit evidence.\n\ n5:25–6:00\nCLOSING PANEL | Industry Career and Opportunities for Studen ts\nGuru Sharan B S — Founder & CEO\, Infi Tech World\nSreejith Kaimal — C3 AI\nChitiz Tayal — Senior Director\, Axtria\nAlok Tibrewala — I ndependent Researcher\n\n6:00–6:05\nClosing Remarks and Acknowledgments\ n\nINTERESTED\nRegistration Link: IEEE Workshop on AI-Native Networking\, Resilience and Forensics - Preliminary Program : vTools Events | vTools IE EE\nSpecial Interest Group: https://ainativenetworking.ieee.org/.\nTo join the SIG send an email to ainativenetwork@ieee.org OR [Join- IEEE AI Nativ e Networking SIG](https://ainativenetworking.ieee.org/contact/#join-sig)\n LinkedIn Group: https://www.linkedin.com/groups/36950078\nTalk/Speaker det ails: https://ainativenetworking.ieee.org/\n\nCo-sponsored by: https://www .linkedin.com/groups/36950078 - AI-Native Networking & Systems Technical C ommunity\n\nBldg: 76\, Carlson Auditorium\, Lomb Memorial Drive\, Rocheste r \, New York\, United States\, 14623\, Virtual: https://events.vtools.iee e.org/m/580732 LOCATION:Bldg: 76\, Carlson Auditorium\, Lomb Memorial Drive\, Rochester \, New York\, United States\, 14623\, Virtual: https://events.vtools.ieee.or g/m/580732 ORGANIZER:nxsvks@rit.edu SEQUENCE:254 SUMMARY:IEEE Workshop on AI-Native Networking\, Resilience and Forensics - Preliminary Program URL;VALUE=URI:https://events.vtools.ieee.org/m/580732 X-ALT-DESC:Description: <br /><p><strong><span style="font-family: 'times n ew roman'\, times\, serif\;">FREE REGISTRATION&nbsp\; OPEN - 23rd OCT 2026 <br></span></strong></p>\n<p>IEEE Workshop on AI-Native Networking\, Resil ience and Forensics<br>DETAILED PRELIMINARY PROGRAM<br>Saturday\, October 24\, 2026 | Carlson Auditorium\, Rochester Institute of Technology<br>8:30 AM&ndash\;6:05 PM &bull\; Technical talks\, panels and student poster awa rds</p>\n<p>Opening &amp\; Session 1 | AI Networking and Infrastructure</p >\n<p><br>8:30&ndash\;8:45 AM<br>Welcome and Introduction &mdash\; Nirmala Shenoy (Rochester Institute of Technology)</p>\n<p>8:50&ndash\;9:20<br>Lo omAI and Autonomous Experiments on the FABRIC Testbed<br>Paul Ruth | RENCI \, University of North Carolina at Chapel Hill<br>Using AI agents to provi sion\, operate\, analyze and reproduce distributed FABRIC experiments\, an d using FABRIC to study agents across real networks.</p>\n<p>9:20&ndash\;9 :40<br>Powering the Next-Gen Campus: The Impact of AI on Infrastructure an d Networking Resources in GCCIS<br>Charles Gruener | Director of Technical Operations\, GCCIS\, Rochester Institute of Technology<br>Campus computin g and networking capacity\, orchestration and resource-management implicat ions of emerging AI workloads.</p>\n<p>9:40&ndash\;10:00<br>Software and N etwork Challenges of Making GPUs Work Together at Scale<br>Deep Shah | Soc ratic AI (formerly Meta and VMware)<br>How distributed AI training require s GPU coordination\, network communication and mechanisms for performance and failure recovery.</p>\n<p>10:00&ndash\;10:20<br>Memory Is the Bottlene ck: Pooling GPUs for One Big Job vs. Slicing One GPU for Many Robots<br>Fr ancis Tyson Thomas | Oishii Farm\; formerly Kroll Government Solutions<br> Compares pooling GPUs across nodes for large jobs with sharing a GPU for r eal-time robotic inference\, including latency and observability tradeoffs .</p>\n<p>10:20&ndash\;10:35<br>MORNING COFFEE BREAK</p>\n<p>Session 2 | A I Optimization and Resilience</p>\n<p><br>10:40&ndash\;11:00<br>Advanced L atent Representations and Computational Optimization<br>Baris Kazar | Infr astructure architect and spatial data scientist\; formerly Oracle America< br>Optimization of embeddings\, fused Triton/CUDA kernels\, vector quantiz ation and approximate nearest-neighbor search for AI systems.</p>\n<p>11:0 0&ndash\;11:20<br>Keeping Distributed AI Inference Reliable Under Overload : Deadline-Aware Backpressure and Adaptive Load Shedding<br>Jayapragash Da kshnamurthy | HealthEquity<br>DA-BALS uses deadline and congestion signals for admission and load shedding in distributed cloud&ndash\;edge AI infer ence.</p>\n<p>11:20&ndash\;11:40<br>When the Retry Is the Outage: Metastab le Failures in Agentic and Distributed AI Systems<br>Abdul Rasheed Shaik | Senior engineering leader\, Fortune 100 firm<br>Simulation-based analysis of retry amplification\, persistent overload and controls such as cancell ation\, retry budgets and admission deadlines. Co-author: Khadarvali Shaik (Fortune 50 healthcare organization)\; Abdul presents.</p>\n<p>11:45&ndas h\;12:25 PM<br>TECHNICAL PANEL 1 | Networking for the AI Era?<br>Vidya Wad kar &mdash\; Verizon<br>Swapna Chimanchodkar &mdash\; Independent Research er<br>Nirmala Shenoy &mdash\; Rochester Institute of Technology<br>Guru Sh aran B S &mdash\; Founder &amp\; CEO\, Infi Tech World</p>\n<p>12:25&ndash \;1:15<br>LUNCH BREAK &amp\; STUDENT POSTER AWARDS</p>\n<p>Session 3 | AI- Native Infrastructure and Operations</p>\n<p><br>1:20&ndash\;1:40 PM<br>Em bedded System Development in Industry: Before and After AI<br>Dr. Varma Ka madi | HCL America (HCL Technologies)<br>Practical changes AI brings to em bedded development\, from interpreting datasheets and writing drivers to d ebugging and testing firmware.</p>\n<p>1:40&ndash\;2:00<br>Bolt-On Intelli gence: Incrementally Engineering an AI Layer for Existing Network Infrastr ucture<br>Gabriel W. Nellis | Morgan Properties\; student\, Rochester Inst itute of Technology<br>A staged approach to network intelligence using man agement interfaces\, shared data models\, deterministic analysis and caref ully governed LLM integration.</p>\n<p>2:00&ndash\;2:20<br>Agentic AI for Predictive Failure Detection and Autonomous Recovery in Cloud-Native Infra structure<br>Rambabu Tangirala | Senior data engineer and independent rese archer<br>Combining observability\, anomaly detection\, retrieval-augmente d generation and AI agents for root-cause analysis and human-approved reco very.</p>\n<p>2:20&ndash\;2:40<br>Scaling Distributed Corporate Infrastruc tures via AI-Native Architectures: An Empirical Framework for Adaptive Res ource Allocation and Deterministic Compliance<br>Guru Sharan B S | Founder &amp\; CEO\, Infi Tech World<br>Enterprise AI architectures using localiz ed vector sharding\, RAG\, multi-agent orchestration and policy-aware hybr id-cloud resource allocation.</p>\n<p>2:45&ndash\;3:20<br>TECHNICAL PANEL 2 | Trustworthy Distributed AI Infrastructure: Security\, Observability an d Forensics<br>Prasanna Kumar &mdash\; M&amp\;T Bank\;<br>Chitiz Tayal &md ash\; Senior Director\, Axtria<br>Venkata Phani Patelkhana &mdash\; Indepe ndent researcher<br>Benjamin Br&uuml\;mm &mdash\; Dalhousie University / O pen Compute Project</p>\n<p>3:20&ndash\;3:35<br>AFTERNOON COFFEE BREAK</p> \n<p>Session 4 | AI Governance\, Security and Forensics</p>\n<p><br>3:40&n dash\;4:00 PM<br>Every Action Should Be Accounted For: Behavioral Forensic s for AI Agents in Data Center Operations<br>Rea Ara | New York University (NYU Tandon)<br>Enumerates agent execution paths and tests deterministic rules against security risks in data-center facilities operations.</p>\n<p >4:00&ndash\;4:20<br>Governing Unreliable AI Services in Regulated Workflo ws<br>Anurag Nedunuri | Independent researcher\; lead solutions architect< br>Auditable asynchronous orchestration of AI requests\, callbacks\, excep tions and human overrides in regulated insurance workflows.</p>\n<p>4:20&n dash\;4:40<br>Beyond Suspicion Scores: Competing-Explanation Inference for Fraud-Shaped Legitimate Activity<br>Ankur Garg | Director of AI and Emerg ing Tech\, Rockland Trust Bank<br>Explaining competing fraud and legitimat e-activity hypotheses to improve transparency and auditability in financia l investigations.</p>\n<p>4:40&ndash\;5:00<br>Enterprise Context Layer: Th e Missing Foundation for Agentic AI<br>Chitiz Tayal | Senior Director\, Ax tria<br>Why governed meaning\, permissions and enterprise context must con nect data platforms with agents to prevent context-driven errors.</p>\n<p> 5:00&ndash\;5:20<br>From Request to Action: Inside the Journey of an AI Ag ent Decision<br>Venkata Phani Patelkhana | Distinguished Engineer and inde pendent researcher<br>Follows user intent through agent identity\, delegat ed authority\, context\, policy checks\, tool execution and audit evidence .</p>\n<p>5:25&ndash\;6:00<br>CLOSING PANEL | Industry Career and Opportun ities for Students<br>Guru Sharan B S &mdash\; Founder &amp\; CEO\, Infi T ech World<br>Sreejith Kaimal &mdash\; C3 AI<br>Chitiz Tayal &mdash\; Senio r Director\, Axtria<br>Alok Tibrewala &mdash\; Independent Researcher</p>\ n<p>6:00&ndash\;6:05<br>Closing Remarks and Acknowledgments</p>\n<p>INTERE STED&nbsp\;<br>Registration Link: IEEE Workshop on AI-Native Networking\, Resilience and Forensics - Preliminary Program : vTools Events | vTools IE EE<br>Special Interest Group: https://ainativenetworking.ieee.org/.<br>To join the SIG send an email to <a href="mailto:ainativenetwork@ieee.org">ai nativenetwork@ieee.org</a>&nbsp\; OR&nbsp\; <a href="https://ainativenetwo rking.ieee.org/contact/#join-sig">Join- IEEE AI Native Networking SIG</a>< br>LinkedIn Group: <a href="https://www.linkedin.com/groups/36950078">http s://www.linkedin.com/groups/36950078</a><br>Talk/Speaker details: https:// ainativenetworking.ieee.org/</p>\n<p><span style="font-family: 'times new roman'\, times\, serif\;">&nbsp\;</span></p>\n<p>&nbsp\;</p> END:VEVENT END:VCALENDAR