Required Skills & Experience
You are a hands-on builder of agentic systems. You demonstrate the complete journey from a nontechnical person’s idea to a working product change, then make it repeatable across real products. Use experiments and evaluations to decide where agents can act independently and where human judgment adds value. Continually evolve the operating model as agent capabilities improve.
-8+ years of software engineering experience, including years of leading engineering teams that build platforms, developer products or complex software systems. -Experience taking an ambiguous product vision from prototype to a dependable, adopted platform in a fast-moving SaaS environment. -Recent, hands-on experience building agentic systems that perform multistep software development work, with depth in orchestration, context engineering, tool use, evaluations and failure recovery. -Strong software architecture and engineering judgment, including how to validate agent-generated changes and operate them safely in production. -Experience designing intuitive workflows for users with different levels of technical expertise, translating intent into executable work. -A product approach to platform leadership: ship early, observe real use, iterate quickly and measure customer outcomes. -Experience integrating agents with existing codebases, cloud environments and delivery systems across multiple technology stacks. -Sound judgment about agent authority, identity, permissions, data access and human oversight, with controls appropriate to change risk. -The ability to lead organizational change and align Product, Design, Engineering, Security and Operations around a new way of creating software.
Nice to Have Skills & Experience
-Experience building software or platforms for regulated SaaS or public-sector environments, including automated assurance evidence. -Experience modernizing acquired or long-lived product portfolios. -Experience measuring and optimizing the reliability, latency and economics of agent workflows.
Job Description
A global government technology organization is seeking a Senior Director, Control Plane to join their team. The Senior Director, Control Plane builds the platform that turns product intent into running software through an agentic product development lifecycle (PDLC). Our ambition is for anyone with a product idea, including a nontechnical product owner, to work with an agent to refine it, inspect working software and direct improvements. Agents carry the work through requirements, design, implementation, validation and deployment, using the Control Plane to access the context, capabilities and controls they need. You own how this model works across the organization: how agents understand our products, coordinate work, make changes, demonstrate correctness and operate within defined authority. You report to the Chief Technology Officer and build a hands-on team with expertise in agent systems, software architecture, product experience and platform engineering. Success means more people can create valuable product changes with less manual coordination and engineering effort. Government agencies rely on these products to serve their residents. This platform must make software creation more accessible while preserving product coherence, security, accessibility and dependable operation. You will:
-Make conversational product creation real. Enable a product owner to describe an idea, refine requirements and acceptance criteria with an agent, review a working preview and direct changes without operating development tools. -Build the system that carries intent through to running software. Orchestrate agents across discovery, design, planning, coding, testing, review and deployment, preserving intent and context throughout. -Give agents the knowledge to make sound product changes: product behavior, customer needs, design systems, architecture, code, dependencies and engineering standards. Keep that context current and scoped to the work. -Design how agents execute and collaborate: task decomposition, tool access, isolated workspaces, durable state, handoffs and recovery. Enable them to resolve routine failures and escalate ambiguity or consequential decisions. -Make quality demonstrable. Build evaluations and automated validation that show whether a change satisfies the intended outcome and meets product, engineering, security and accessibility standards, with traceable evidence. -Establish the path to greater autonomy. Begin with reviewable PRs and human release decisions; expand to autonomous completion and deployment where demonstrated reliability, change risk and policy permit. Build in permissions, observability, intervention and rollback. -Make the model work across our existing portfolio. Identify what prevents agents from understanding, changing or validating each product, and partner with product teams to remove those barriers. -Run the Control Plane as a product. Ship useful capabilities quickly, learn from technical and nontechnical users, and expand adoption based on demonstrated outcomes. -Build a system that improves through use. Turn user feedback, agent failures, review findings and production outcomes into better context, workflows and evaluations. -Build and lead the team that delivers this platform. Partner with Product, Design, Engineering, Security and Operations to encode standards and regulatory obligations, including FedRAMP, into how agents work. -Measure time from idea to validated running software, successful completion of agent workflows, human effort per change, adoption by nontechnical users, production quality and cost per successful outcome.