Erik Koci engineers cloud-native infrastructure and automation tools at Amadeus. He focuses on bringing artificial intelligence directly into operational systems without the usual complexity.
Working primarily with Kubernetes, Nomad, Go, and Python, he builds active bridges between LLMs and real-world platforms. His recent work centers on the Model Context Protocol (MCP). He built a Go-based MCP server that lets AI models interface directly with Nomad clusters to read logs, collect metrics, and run jobs.
Erik is a regular speaker in the HashiCorp ecosystem, where he shares practical methods for keeping infrastructure scalable and simple.