Mycelium Cloud

A managed Kubernetes service running on a distributed compute platform, where the hardware underneath a cluster is not centrally owned and fails on its own schedule.

Skills

Go · Kubernetes · k3s · CRDs · PostgreSQL

Contributions

• Built the cluster deployer engine for multi-master k3s clusters, with persistent workflow state and exponential-backoff retries that resume cleanly through partial failures.
• Designed the rollback workflow that unwinds failed deployments step by step in reverse, without operator intervention.
• Extended Kubernetes with custom resources and controllers that auto-reconcile gateway and ingress routing.
• Shipped a real-time notification system streaming workflow progress to the dashboard over SSE.