Job Description

What is the Opportunity?

We're building the agentic platform that will let every development team at RBC execute AI driven SDLC — our spec-driven, agent-assisted software delivery methodology — at scale. This is a hands-on platform engineering role inside DevOps: you won't be running one team's pipeline, you'll be building the harness, skills, and agents that every team's development workflow runs on.

If you're the kind of engineer who has already wired tools like Claude Code into your own workflow with custom skills and MCP servers to make yourself dramatically more productive, and you want to turn that instinct into a platform used across an entire bank, this is that role.

What will you do?

• Design and build agents across the full software delivery lifecycle: specification creation, technical spec generation, build-and-test agents, and operational agents that automate deployment, release, PIV (post-implementation verification), and post-deployment tasks
• Build these agents on coding harnesses via custom skills and tools (including MCP servers), on managed agent platforms such as Devin, and as native agents using frameworks like LangChain and LangGraph — choosing the right substrate for each use case
• Implement and extend spec-driven development workflows (e.g., specify → clarify → plan → tasks → analyze → implement, in the style of GitHub spec-kit / SpecKit) as the backbone of the platform's golden path
• Build the technical plumbing that enforces AI-DLC's gates (e.g., Spec Ready, PR Ready) as machine-checked controls rather than advisory guidance
• Continuously push what frontier models and coding harnesses like Claude Code can automate — both at a personal productivity level and as reusable capability that scales across other development teams
• Partner with DevOps, security, and platform teams to get new agents and skills safely into production and into the shared skill/tool catalog
• Evaluate emerging agentic AI platforms, frameworks, and design patterns, and make build-vs-adopt recommendations for the platform roadmap
• Regularly demonstrate what you've built to other development teams — explaining not just what an agent does, but how it works — to build trust and drive adoption of the platform
• Act as a hands-on technical resource and champion for teams adopting AI-DLC, helping them get unblocked on the platform

What do you need to succeed?

Must-have:

• Strong software development background (Python and/or TypeScript/JavaScript) with the ability to build and operate production-grade services, not just prototypes
• Hands-on experience building agent-based solutions — e.g., with LangChain, LangGraph, or comparable agent frameworks – even within coding harnesses like Claude Code.
• Practical experience with MCP (Model Context Protocol) servers and building custom skills/tools for coding harnesses such as Claude Code
• Demonstrated personal fluency with frontier coding assistants (e.g., Claude Code) for real engineering work, including multi-agent workflows
• Experience with CI/CD pipelines and DevOps tooling (e.g., GitHub Actions, Jenkins, Docker/Kubernetes) — you understand what deployment, release, and post-deployment automation actually needs to do
• Familiarity with wider software development lifecycle, including product discovery, stakeholder requirements gathering, design and architecture
• Strong communicator, comfortable presenting and demoing technical work to other development teams and translating platform capabilities into terms application teams can act on
• Strong software architecture and system design skills; comfortable making platform-level tradeoffs that will be consumed by many other teams