Principal, AI Agent Control & Governance

At ATB, we exist to make it possible for our clients, team members and communities. Our purpose is more than aspirational. It's a real commitment we live every day through our values (what we call the ATB ID).

Life at ATB is about more than work. In fact, we’re consistently recognized as one of Canada's top employers thanks to our high-care, high-performance culture, upheld by the three commitments we make to our team members:

Impact in action: No matter where you are in the organization, you're empowered to make an impact in the lives of our clients and communities. 

Thrive together: We want you—the unique, authentic you—to feel safe and celebrated at work. We're on a continuous journey to build the most flexible and inclusive programs. 

Ready for tomorrow: We want to enable your success through interesting and challenging work, performance enablement, and learning and development.

About the role

This position has a hybrid work arrangement. On average, this means 1-2 day(s) on-site per week. Please note that this is subject to change due to business priorities.

As ATB’s next Principal, AI Agent Control & Governance you will hold end-to-end operational and technical accountability for how AI tools, frontier models, and autonomous agents operate across ATB. Positioned at the intersection of AI platform engineering, autonomous system deployment, and enterprise risk oversight, this role bridges technical development, solution sponsors, and enterprise control functions.

This role encompasses two connected mandates:

1. AI Operations & Platform Administration: Leading the operationalization, core configuration, licensing, FinOps cost optimization, and access controls for enterprise AI platforms and frontier models (e.g., Google Gemini, Anthropic Claude, OpenAI Enterprise, Microsoft Copilot).

2. AI Agent Governance & Lifecycle Management: Owning the authoritative enterprise registry for all AI agents (in intake, development, or production), maintaining runtime observability, and detecting "agents in the wild" or emerging model/agent drift risks.

The Principal will architect, execute, and scale operational AI solutions and agent governance frameworks across ATB's enterprise AI platform infrastructure. The Principal, AI Agent Control & Governance operationalizes governance requirements directly into technical workflows, MLOps pipelines, and agent execution environments. This ensures innovation velocity is maximized while maintaining audit-ready alignment with OSFI regulatory expectations (Guideline E-23 Model Risk Management, E-21 Operational Risk), enterprise risk appetites, and operational safety guardrails.

Accountabilities

• Enterprise Model & Platform Administration: Lead the operational deployment, lifecycle management, licensing agreements, and core configurations for frontier models (Google Gemini, Anthropic Claude, etc.) and enterprise productivity tools.

• IAM & Access Control: Partner directly with the Identity and Access Management (IAM) team to implement and manage granular Role-Based Access Control (RBAC), tiered permissions, and identity controls across platforms.

• FinOps & Cost Optimization: Track, analyze, and optimize platform usage metrics and cloud/AI spend to support executive decision-making on licensing tiers and resource allocation.

• Service Level Operations (SLO/SLA): Act as the primary technical point of contact for platform configurations, API integrations, and service resolutions, maintaining adherence to defined operational SLAs and SLOs.

• Observability & Application Health: Establish monitoring systems to track AI application health using observability frameworks (logging, evaluation operations metrics, LLM latency, and performance metrics) to resolve issues proactively.

• Security Guardrails & AppSec Collaboration: In collaboration with Application Security, implement platform-level AI security guardrails and conduct risk assessments to guarantee compliance, privacy, and safety.

• Cross-Functional Model Harnessing: Partner with Data Scientists, Software Engineers, and product managers to define model harness operational requirements and track business outcomes for AI products in production.

• Enterprise Asset Registry: Own and maintain the authoritative, audit-ready AI Asset Registry capturing all models, third-party tools, API dependencies, fine-tuned foundational models, and agentic workflows across ATB.

• Discovery of Unsanctioned Systems: Identify and evaluate tooling to maintain visibility into "agents in the wild" operating outside formal intake channels.

• Agent Architecture & Guardrails: Collaborate with AI Engineering, Security, Data, and Architecture teams to design agent orchestration layers and implement operational governance guardrails.

• Human-in-the-Loop (HITL) Controls: Define operational protocols specifying human intervention thresholds, supervisory checkpoints, and fallback mechanisms based on agent autonomy and risk impact.

• Behavior & Drift Monitoring: Establish continuous runtime protocols to monitor agent behavior for prompt injection vulnerabilities, hallucination, non-deterministic execution, and model drift.

• Post-Deployment Auditability: Ensure comprehensive activity logging, reasoning trace retention, and audit trails for all automated decisions driven by agentic workflows.

• Regulatory Standard Mapping: Recognize and understand regulatory expectations in collaboration with internal frameworks, policies, and controls.

• Control Partner Collaboration: Work with AI Risk & Compliance, Privacy, Legal, and Data Governance to streamline risk reviews without directly inheriting compliance decision ownership.

• Incident Containment & Exception Handling: Lead operational protocols for AI performance incidents, model failure containment, safety guardrail breaches, and documented governance exceptions.

Skills, experience & requirements

• Education: Bachelor’s or Master's degree in Computer Science, Information Systems, Software Engineering, Risk Management, or a related technical discipline.

• Overall Experience: 5+ years of progressive experience in AI/ML platform administration, IT enterprise operations, system architecture, or technical governance within regulated environments.

• Big Data & ML Deployment: 5+ years of direct experience in big data projects, specifically supporting and deploying Machine Learning models into production.

• Systems Infrastructure: 5+ years of experience in enterprise cloud infrastructure and information systems environments.

• Agile & Delivery: 3+ years of experience working with Agile methodologies and enterprise collaboration tools (Jira, Confluence).

• Frontier LLMs & Platforms: Deep hands-on experience administering enterprise LLM platforms (Google Gemini Enterprise, Anthropic Claude, OpenAI Enterprise, Microsoft Copilot), including API integrations, RBAC, and cost management.

• Programming Proficiency: 5+ years of experience writing code in languages such as Python, R, Java, JavaScript, or Bash script.

• MLOps & Pipeline Orchestration: Advanced knowledge of Machine Learning Operations (MLOps), pipeline orchestration, and Software Development Life Cycle (SDLC) / DevOps practices.

• Cloud Infrastructure: Practical experience implementing MLOps pipelines on major cloud platforms (Google Cloud Vertex AI, AWS SageMaker, Azure OpenAI).

• MLOps & Observability Frameworks: Familiarity with orchestration tools (Kubeflow, MLflow, DataRobot, Airflow) and LLM observability mechanisms.

• Agentic Frameworks: Hands-on experience or familiarity with agent development SDKs and orchestration frameworks (e.g., ADK, LangGraph, Claude Agent SDK).

• Containerization: Practical experience with container technologies including Docker, Kubernetes, and OpenShift.

• Regulatory Alignment: Strong familiarity with financial industry standards (OSFI E-23, OSFI E-21) and technology risk frameworks (NIST AI RMF).

• Executive Communication: Exceptional ability to translate complex technical architectures and regulatory directives into clear, plain-language summaries for executive stakeholders.

• Stakeholder Facilitation: Proven ability to bring together technical developers, data scientists, security experts, legal counsel, and business sponsors into operational alignment.

Don’t meet all the requirements on the list?
A resume only goes so far in expressing who you are and the unique perspectives you bring. If you believe your skills and experience align with the role—but you might not check all the boxes—we want to hear from you. We encourage candidates from all work backgrounds, equity-seeking communities and experience levels to apply. If you’re seeking a career where your drive, perspective and growth are celebrated, we want to hear from you.

We’re dedicated to building a workforce reflective of those within our communities, and a culture where our team members are equipped with what they need to succeed—their way. Part of creating an inclusive workplace is recognizing our role in advancing Truth and Reconciliation. We are committed to meeting and exceeding the standards set out in the Partnership Accreditation in Indigenous Relations program created by the Canadian Council for Indigenous Business. 

What happens next?
If you are shortlisted, you’ll hear from us after the posting closes. Check out our How We Hire page to learn more about our hiring process. If you need any accommodations throughout this process, please let us know at talentteam@atb.com

Stay in touch
We’re excited that you’re interested in a career with ATB. Follow us on LinkedIn, Facebook and Instagram to learn more about what our team is up to.