apiVersion: homayoon.dev/v1
kind: Engineer
metadata:
name: homayoon-alimohammadi
labels:
role: senior-backend-engineer
team: canonical/kubernetes
spec:
focus:
- distributed systems and microservices that stay correct under load
- Kubernetes internals, operators, and cluster lifecycle
- platform work whose users are other engineers
languages: [go, python, sql, bash]
currentlyBuilding:
- k8sd — the clustering daemon behind every Canonical Kubernetes cluster
- Cluster API bootstrap & control-plane providers (AWS, GCP, MAAS, bare metal)
- an upgrade controller that converges a cluster from one declarative apply
status:
currentLocation: Dubai, UAE
openTo: senior backend / platform roles across the EU / relocation
phase: Running| nodes scale ceiling established for Canonical Kubernetes |
messages / day event-driven notification platform over Kafka |
users served by the services I owned at Divar |
| requests / second sustained on production Go services |
availability SLO held through a 4-release zero-downtime redesign |
p99 latency cut and 60% less kube-apiserver load at scale |
| microcni | A minimal container network interface plugin — CNI from first principles |
| k8s-doctor | Answers Kubernetes questions from augmented upstream docs, then reproduces issues and verifies fixes on its own managed cluster |
| k8s-mcp-server | MCP server exposing a Kubernetes cluster to LLM agents |
| rag-mcp | MCP server for retrieval over ingested documents and sources |
| k8sutils | Node draining utility for safe cluster maintenance |
| gosync | Go synchronization primitives, rebuilt to understand them properly |
gRPC · REST · GraphQL · Protobuf · etcd · ClickHouse · OpenTelemetry · controller-runtime · Cluster API · pprof · LLM agents & MCP
I write about Go, distributed systems, and Kubernetes internals — and I've authored 10+ design documents through to production.
Still figuring it out, in public.