As a Principal Software Engineer, you will drive the evolution of the GCP Hosted Control Planes (HCP) platform, enabling managed OpenShift clusters on Google Cloud through HyperShift. This role sits at the forefront of a transformative engineering model—shifting from traditional development to an agent-first paradigm where AI-powered coding agents play a central role in building and maintaining systems.
Architecture & System Design
You will define the foundational architecture for a globally distributed, multi-tenant control plane, ensuring scalability across regions, strong isolation guarantees, and resilient automated operations. Your designs will shape how clusters are provisioned, upgraded, and monitored at scale, with a focus on long-term maintainability and operational excellence.
Harness Engineering Leadership
You will build and refine the infrastructure that guides both human and agent contributions—establishing architectural constraints, custom linters, structural tests, and CI enforcement gates. This includes designing feedback mechanisms that allow agents to operate safely and effectively, while preventing degradation and technical drift over time.
Knowledge & Documentation Architecture
You will structure and maintain a living documentation system that serves as the shared source of truth for engineers and AI agents alike. This includes curating design documents, architecture decision records, and operational runbooks, with AGENTS.md acting as the central index that connects agents to context.
Technical Mentorship & Collaboration
You will guide senior engineers in harness engineering practices—crafting precise specifications, designing agent-compatible interfaces, and reviewing machine-generated code with critical rigor. You will lead architectural discussions across internal teams, open source communities, and cloud platform integrations, influencing design decisions beyond your immediate scope.
Operational & Production Excellence
You will define quality standards, testing strategies, and readiness criteria for features developed by agents, ensuring they meet production reliability expectations. As a technical escalation point, you will diagnose and resolve complex customer issues and participate in on-call rotations to support managed services.
Community & Technical Influence
You will maintain an active presence in the Kubernetes, OpenShift, and GCP ecosystems, contributing to upstream projects and advancing best practices in cloud-native infrastructure. Your work will help shape the future of AI-augmented software development in large-scale distributed systems.
Required Expertise
- 10+ years of software engineering experience with strong proficiency in Go
- Deep understanding of Kubernetes internals, including controllers, operators, and cluster lifecycle management
- Proven experience designing and maintaining large-scale distributed systems in production
- Hands-on experience with a major public cloud provider—GCP preferred
- Experience defining architectural standards and coding conventions across teams
- Strong technical writing skills, with an emphasis on structured, executable documentation
- Familiarity with AI-assisted development workflows and the challenges of machine-generated code
- Ability to lead and influence across organizational boundaries without direct authority
Preferred Qualifications
- Experience with GKE, GCP networking, IAM, and Workload Identity Federation
- Background in HyperShift, Cluster API, or multi-tenant Kubernetes platforms
- Development of custom linters, static analysis tools, or architectural test suites
- Proficiency with infrastructure-as-code and GitOps tools such as Terraform, Tekton, and ArgoCD
- Experience with large-scale observability using Prometheus, Google Managed Prometheus, and distributed tracing
- Contributions to open source, especially in the Kubernetes ecosystem
- Operational experience with SLA-bound managed services
- Design of documentation or context systems for AI/LLM tooling
- Knowledge of harness engineering principles—entropy control, constraint design, feedback loops
