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:20260419T200237Z
UID:69343633-2E26-4648-B8D8-32034EA0EAE9
DTSTART;TZID=America/New_York:20260901T180000
DTEND;TZID=America/New_York:20260901T200000
DESCRIPTION:Technical Presentation and Dinner presented by the IEEE Susqueh
 anna Section\nIn highly regulated industries\, the "busy work" of security
  patching often lags due to strict data privacy requirements that prevent 
 the use of cloud-based AI. This session demonstrates how to bridge that ga
 p by building a privacy-first DevSecOps agent that operates entirely withi
 n your own firewall.\n\nWe will explore an architecture that leverages Lla
 ma 3.1 via Ollama and the Model Context Protocol (MCP) to create a structu
 red OODA loop (Observe\, Orient\, Decide\, Act). By connecting Trivy for v
 ulnerability scanning directly to GitHub for automated remediation\, we sh
 ow how an agent can autonomously draft patches while maintaining a strict 
 Human-in-the-Loop (HITL) framework for final verification.\n\nKey Takeaway
 s:\n- Local-First Architecture: How to run a CPU-friendly\, production-rea
 dy AI agent on standard cloud instances or local hardware.\n- Tool Orchest
 ration: Using MCP to bridge the gap between security scanners and version 
 control systems.\n- Correctness & Safety: Implementing HITL checkpoints to
  ensure AI-generated patches are reviewed for correctness before deploymen
 t.\n\nFood and drink will be provided free of charge to attendees.\n\nCo-s
 ponsored by: Penn State Harrisburg School of Science\, Engineering\, and T
 echnology\, Electrical Engineering Department\n\nSpeaker(s): Pathmanaban P
 alsamy\, \n\nAgenda: \nDinner: 6:00 - 7:00 PM\n\nTechnical Presentation: 7
 :00 - 8:00 PM\n\nWrap-Up: 8:00 - 8:15 PM\n\nRoom: 101 (Morrison Gallery)\,
  Bldg: Madlyn L Hanes Library (Building D)\, Penn State Harrisburg\, 777 W
 est Harrisburg Pike\, Harrisburg\, Pennsylvania\, United States\, 17507\, 
 Virtual: https://events.vtools.ieee.org/m/541535
LOCATION:Room: 101 (Morrison Gallery)\, Bldg: Madlyn L Hanes Library (Build
 ing D)\, Penn State Harrisburg\, 777 West Harrisburg Pike\, Harrisburg\, P
 ennsylvania\, United States\, 17507\, Virtual: https://events.vtools.ieee.
 org/m/541535
ORGANIZER:jack.burbank@ieee.org
SEQUENCE:5
SUMMARY:Privacy-First DevSecOps: Automating Vulnerability Remediation with 
 Local AI Agents
URL;VALUE=URI:https://events.vtools.ieee.org/m/541535
X-ALT-DESC:Description: <br /><p>Technical Presentation and Dinner presente
 d by the IEEE Susquehanna Section</p>\n<div>In highly regulated industries
 \, the "busy work" of security patching often lags due to strict data priv
 acy requirements that prevent the use of cloud-based AI. This session demo
 nstrates how to bridge that gap by building a privacy-first DevSecOps agen
 t that operates entirely within your own firewall.</div>\n<div>&nbsp\;</di
 v>\n<div>We will explore an architecture that leverages Llama 3.1 via Olla
 ma and the Model Context Protocol (MCP) to create a structured OODA loop (
 Observe\, Orient\, Decide\, Act). By connecting Trivy for vulnerability sc
 anning directly to GitHub for automated remediation\, we show how an agent
  can autonomously draft patches while maintaining a strict Human-in-the-Lo
 op (HITL) framework for final verification.</div>\n<div>&nbsp\;</div>\n<di
 v>\n<div><strong>Key Takeaways:</strong></div>\n<div>- Local-First Archite
 cture: How to run a CPU-friendly\, production-ready AI agent on standard c
 loud instances or local hardware.</div>\n<div>- Tool Orchestration: Using 
 MCP to bridge the gap between security scanners and version control system
 s.</div>\n<div>- Correctness &amp\; Safety: Implementing HITL checkpoints 
 to ensure AI-generated patches are reviewed for correctness before deploym
 ent.</div>\n<div>&nbsp\;</div>\n<div>Food and drink will be provided free 
 of charge to attendees.</div>\n</div><br /><br />Agenda: <br /><p>Dinner: 
 6:00 - 7:00 PM&nbsp\;</p>\n<p>Technical Presentation: 7:00 - 8:00 PM</p>\n
 <p>Wrap-Up: 8:00 - 8:15 PM</p>
END:VEVENT
END:VCALENDAR