Build and extend Kubernetes orchestration for AI inference: implement CRDs/controllers in Go, design high-performance ingress/load-balancing, implement GPU-aware autoscaling (HPA/VPA), tune scheduler/resource management, and ensure HA for platform components.

Design and operate GPU/ML CI/CD and release pipelines (GitHub Actions, self-hosted H100/A100 runners). Implement packaging, multi-arch containers, continuous benchmarking/performance gates, contributor-friendly CI, and infrastructure-as-code with security and autoscaling.