Principal AI Engineer, AI DLC Platform

What is the opportunity?

The AI-DLC Platform embeds governed, agentic AI across RBC's software development lifecycle - skills, commands and guardrails spanning discovery, build/test/security and release, now running in real repositories. The next step is to take it from a governed foundation (human-approved, A0-A2) to hybrid skills and agents, and ultimately a software factory, with autonomy added only where evidence supports it.

As Principal AI Engineer you will own the technical direction of that journey. You will define the platform architecture, the production boundary between code that runs in production and agents that run in production, and the standards, pipelines and runtime that make AI-assisted delivery auditable, reproducible and safe at enterprise scale. You will provide technical leadership across multiple squads and domains, advise senior leadership on AI-DLC architecture and engineering, and be accountable for the adoption and measurable business impact of the platform.

This role suits a technical leader who combines deep agentic-AI and platform engineering skill with the judgment to work with risk, governance and engineering partners and turn policy into something that can be tested.

What will you do?

• Own the AI-DLC Platform architecture (skills, commands, agents, context layer, MCP services, governance) and define the 12-18 month technical roadmap from governed foundation through hybrid skills/agents to software factory.
• Define the autonomy model (A0 and beyond) with evidence, controls and approvals required before capabilities earn more autonomy.
• Make high-impact build-vs-buy and vendor decisions for AI tooling, models and infrastructure; evaluate open-source vs. proprietary trade-offs.
• Own the promotion pipeline from Harness validation through Helios deployment; ensure AI-generated code is auditable, reproducible, explainable and traceable to specs and prompts.
• Define and enforce the boundary between AI-generated code and sanctioned agents in production; codify governance policy with Risk and Engineering teams.
• Establish coding standards and spec-driven development practices (spec-kit, OpenSpec) so AI output is structurally sound before human review.
• Design and build multi-agent workflows with orchestration frameworks (LangGraph, CrewAI, AutoGen) including tool use, memory, planning, guardrails and human-in-the-loop patterns.
• Design tool contracts, APIs (FastAPI/gRPC) and MCP services so skills and agents are versioned and governed; build and operate the runtime with containerization, orchestration, scaling and observability. Engineer reliability into production systems through circuit breakers, graceful degradation, rollback and zero-downtime deployment; establish LLMOps including evaluation, governance, RAG strategy and cost/latency optimization.
• Build monitoring and alerting that distinguishes bad models, code and platform issues; use AI to auto-generate test cases and validation checks; define the evaluation framework for certifying skills and agents.
• Implement access controls, audit trails and human-in-the-loop validation; set SLAs and reliability targets; own technical risk assessment for AI initiatives.
• Be accountable for platform adoption, outcomes and ROI; define and track KPIs (cycle time, quality, security, usage) and report to senior leadership.
• Advise and present to executive steering committees on AI-DLC architecture, direction, progress and impact; partner with cohort teams to drive adoption.
• Lead architecture design, code reviews and technical discussions; mentor engineers, participate in hiring and define competency frameworks for AI engineers.
• Document and share platform knowledge; represent the platform in internal and external forums.

What do you need to succeed?

Must-have

• Agentic AI architecture: Proven expertise designing and orchestrating multi-step autonomous agents with tool use, memory, planning, guardrails and human-in-the-loop patterns; hands-on proficiency with LangGraph, CrewAI or AutoGen in production environments.
• LLM engineering and LLMOps: Deep competency in prompt engineering at scale, RAG and context management, evaluation frameworks, cost/latency optimization, and model lifecycle management in productio.
• Tool and data integration: Strong capability designing tool contracts, APIs (FastAPI/gRPC) and MCP connections
• CI/CD and infrastructure: Demonstrated mastery of CI/CD pipelines, container orchestration (Docker/Kubernetes) and infrastructure-as-code; proven work with enterprise platforms (e.g. OpenShift, Kafka, Elasticsearch, Apigee) and cloud providers (AWS, Azure, GCP).
• Spec-driven development: Ability to translate user stories and requirements into precise specifications for automated execution; familiarity with frameworks like spec-kit or OpenSpec.
• Automated testing and QA: Proficiency using AI to generate tests, analyze logs and run validation checks; background building scalable test automation.
• Governance and security: Demon