BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//pretalx//pretalx.devconf.info//devconf-cz-2026//talk//EYPGF7
BEGIN:VTIMEZONE
TZID:CET
BEGIN:STANDARD
DTSTART:20001029T040000
RRULE:FREQ=YEARLY;BYDAY=-1SU;BYMONTH=10
TZNAME:CET
TZOFFSETFROM:+0200
TZOFFSETTO:+0100
END:STANDARD
BEGIN:DAYLIGHT
DTSTART:20000326T030000
RRULE:FREQ=YEARLY;BYDAY=-1SU;BYMONTH=3
TZNAME:CEST
TZOFFSETFROM:+0100
TZOFFSETTO:+0200
END:DAYLIGHT
END:VTIMEZONE
BEGIN:VEVENT
UID:pretalx-devconf-cz-2026-EYPGF7@pretalx.devconf.info
DTSTART;TZID=CET:20260618T110000
DTEND;TZID=CET:20260618T113500
DESCRIPTION:Why do AI agents remain fragile and tightly coupled? Most produ
 ction implementations lack a standardized way to interact with the OS and 
 external tools. This session explores building a modular\, cloud-native ag
 ent stack using Model Context Protocol (MCP)—an open standard acting as 
 a "USB-C port" for AI interoperability.\n\nIn this talk\, we’ll share ho
 w we integrated MCP into an open AI agent architecture using:\n1) vLLM for
  high-performance model inference\n2) LlamaStack as the agentic orchestrat
 ion framework\n3) MCP for standardized tool invocation and data flow\n4) K
 ubernetes for managing distributed agent workloads\n\nWe’ll demo the sys
 tem and share lessons on scaling these components in a cloud-native enviro
 nment like Kubernetes. You’ll learn how MCP addresses the interoperabili
 ty gap and helps avoid vendor lock-in. Attendees will walk away with a roa
 dmap for building modular\, vendor-neutral AI applications that are firmly
  in the hands of the community.
DTSTAMP:20260430T125800Z
LOCATION:D105 (capacity 300)
SUMMARY:Decoupling AI Agents: Building a modular stack with MCP and Kuberne
 tes - Hema Veeradhi\, Yu An
URL:https://pretalx.devconf.info/devconf-cz-2026/talk/EYPGF7/
END:VEVENT
END:VCALENDAR