AI Architect - Developer Enablement

We need a proper builder to own how AI is used to develop software across the organisation.

Not an AI strategy role. Not a chatbot role. This is about building the developer AI ecosystem: the model layer, coding agents, agentic loops, developer tooling, specs, repositories, CI pipelines, testing and the standards that let engineers use AI properly at scale.

You will work out what actually works in practice. Which models are good for planning, coding, debugging, review and testing? How should developers use K3, GLM and other frontier/open models? Where do agentic loops help, where do they go off the rails, and how do we give them the right context, tools, permissions and controls?

You will define and build:

• AI coding workflows from specification through build, test, review and deployment
• Agentic coding loops: plan, execute, validate, fix, retry and escalate
• Shared context for agents: codebase knowledge, architecture, standards, tickets, APIs and documentation
• Tool integrations around source control, IDEs, CI/CD, test suites, security scanning and developer environments
• Reusable skills, prompts, agent patterns and templates for different engineering jobs
• Clear standards for vibe coding: when it is useful, what must be tested, reviewed and evidenced before code is merged
• Model selection, routing, cost, performance, privacy and access controls
• Evaluation of AI-generated code: correctness, security, maintainability, test quality and production outcomes

You need to be a seriously capable software engineer with a strong understanding of the fast-moving AI coding ecosystem. You should know how coding agents, model providers, MCP/tool calling, context engineering, local and hosted models, repositories and development platforms fit together.

If you enjoy experimenting hard, separating hype from what genuinely improves engineering, and then building the platform that makes it usable for everyone else, this is the role.