Director, Agentic Engineering

About The Role

Fifth Third is standing up a new Agentic Engineering team within Technology. We are building autonomous agents, multi‑agent architectures, and the platform capabilities that make them safe, observable, and production‑ready inside a regulated financial institution. The Director of Agentic Engineering will build and lead this team from the ground up while remaining deeply hands‑on in architecture and code.

• Design and build autonomous agent and multi‑agent systems for high‑value bank workflows.
• Define and implement tool‑use patterns, agent orchestration, and Model Context Protocol (MCP) integrations that enable agents to operate safely across enterprise systems.
• Write and review production code in critical system components — this role leads by example.
• Establish approved architectural patterns for agentic development, enabling teams across the organization to build and explore AI solutions securely.

• Own the guardrail enforcement layer for agentic systems — designing, implementing, and continuously improving safety controls for autonomous agent behavior.
• Partner with risk, compliance, legal, and information security to ensure agent systems meet regulatory and enterprise governance requirements.
• Build evaluation, observability, and debugging capabilities that provide transparency into agent decision‑making, tool usage, and failure modes.
• Define agent governance standards that align with enterprise risk appetite and regulatory expectations.

• Build and maintain APIs, SDKs, and developer tooling that accelerate safe AI development across the enterprise.
• Define and secure approved patterns and reference architectures that lower the barrier to responsible agentic development.
• Create reusable frameworks, templates, and playbooks that convert repeatable patterns into organizational assets.

• Recruit, hire, and build the Agentic Engineering teams.
• Mentor and coach engineers across career levels — developing individual contributors into technical leaders who are both technically excellent and deeply user‑oriented.
• Partner with senior business and technology leaders to identify where autonomous agents can deliver measurable business impact.
• Establish delivery patterns for safe experimentation, piloting, production release, measurement, and ongoing support.
• Maintain a practical view of emerging AI capabilities and where they can be responsibly applied inside a regulated financial institution.

• Experience leading highly technical engineering teams responsible for production systems at enterprise scale.
• Hands‑on experience building with AI agent frameworks (e.g., Lang Chain, Lang Graph, Deep Agents) and deploying agentic systems to production.
• Direct experience building and deploying MCP servers and tool‑use integrations.
• Strong background in cloud architecture and system design on AWS (Bedrock, Lambda, Fargate, IAM, networking).
• Production coding proficiency across multiple languages (Python, Type Script/Node.js preferred).
• Extensive experience in automation, infrastructure‑as‑code (Terraform), and platform engineering.
• Proven track record delivering world‑class AI‑based architecture aligned with successful business outcomes.
• Background in security, enterprise architecture, or systems engineering sufficient to partner credibly with risk and compliance teams.
• Experience working in banking, financial services, or another regulated industry.
• Ability to operate as a player‑coach — setting technical direction, reviewing code, and developing people simultaneously.
• Strong product judgment and the ability to separate high‑value agent applications from novelty.
• Clear communication style, especially when translating between business, engineering, and risk stakeholders.

• Experience with agent‑to‑agent (A2A) communication patterns and multi‑agent coordination.
• Background in developer experience, internal platform engineering, or API/SDK design.
• Experience with agent evaluation frameworks, LLM observability tooling, and production debugging of non‑deterministic…