Register: MCP Academy PNW · Luma
Date and location: Wednesday, October 29 - Bellevue, Washington
Talk
Using a real-world incident response automation server as example, in just 15 minutes I’ll show you how to flip your thinking from a task-oriented to a goal-oriented mindset. In this talk, I will walk you through a goal-driven approach to building MCP servers that enable AI agents to pursue complex objectives independently while maintaining safety and control.
Key Takeaways
Attendees will leave with actionable knowledge to:
- Design MCP servers that enable autonomous agent behavior while maintaining human oversight
- Map complex business workflows into phase-based architectures suitable for AI orchestration
- Implement outcome-focused tools that provide rich context and control to AI agents
- Build safety mechanisms that prevent runaway automation while preserving efficiency
- Apply these patterns to diverse domains beyond incident response (content publishing, deployment pipelines, customer support workflows)
Target Audience
This talk is for experienced software engineers, architects, and DevOps professionals who are interested in the future of AI and automation. Attendees should have an understanding of software design principles and be familiar with building and consuming APIs. No prior experience with MCP is required.
Speaker
I am a a senior software engineer with experience in Go, distributed systems, and open source, having served as a core maintainer for Cloud Native projects including Velero and Knative. My unique perspective combines engineering experience with recent explorations at the intersection of AI, biology, and system design.