Principal Software Engineer, Agentic Engineering

• Lead and grow a team of software engineers building production agents - setting direction, reviewing designs, mentoring, and unblocking.

• Define the team’s engineering standards - agent patterns, tool integration, testing, code quality, observability, and review practices that hold the bar high.

• Architect the agent platform - the shared services and protocols (e.g., agent-to-agent messaging, tool servers) that let agents discover and compose with one another reliably.

• Stay hands-on - write code, prototype hard problems, and pair with engineers on the toughest parts of the system.

• Own the quality and safety of agent behavior - design how the team verifies that agents stay accurate, cost-efficient, and within guardrails over time.

• Translate vision into a roadmap - break an ambitious platform vision into shippable increments the team can execute and be proud of.

• Represent the team to leadership - communicate architecture, trade-offs, and progress clearly to engineering and product leadership.

What Differentiates This Role

This role asks you to work with engineers to build autonomous systems that reason - agents that make decisions, call tools, and message one another.

• You’re defining how agents are built across the team - the shared patterns, base abstractions, and “paved road” that make new agents fast to build and safe to run.

• You’re responsible for how we know agents are behaving - observability, evaluation, and continuous governance, not a one-time check.

• You’re growing a team that can reason about non-deterministic systems - where correctness, cost, and safety all have to be designed in, not bolted on.

We are looking for a Principal Software Engineer to join our Agentic Engineering team. You will build a platform where specialized AI agents compose across technical domains - knowledge and policy, change/ITSM, network security, cloud, and source code. Instead of an engineer manually stitching together answers from a half-dozen tools, agents coordinate with each other to do the work.This is a hands-on technical leadership role that requires deep proficiency in Python and TypeScript, a track record of designing systems that operate reliably at organizational scale, and the ability to set standards that shape how enforcement is practiced across teams., + 8-12+ years of software engineering experience with deep proficiency in Python and TypeScript.

• A track record of technically driving teams of engineers - mentoring, setting standards, and shipping complex systems together.

• Experience building agentic or multi-step automated systems - tool orchestration, LLM reasoning loops, or agent frameworks.

• The ability to set architectural direction and reusable patterns that other engineers adopt.

• Experience designing for non-deterministic behavior, edge cases, and failure - and building the testing and observability to catch them.

Nice to Have

• Experience building LLM safety, evaluation, or governance systems end-to-end.

• Retrieval / hybrid search over enterprise knowledge.

• Background in platform engineering or developer tooling at enterprise scale.

• Terraform and infrastructure-as-code experience.

• Open-source contributions on public GitHub.

At Fifth Third, we understand the importance of recognizing our employees for the role they play in improving the lives of our customers, communities and each other. Our Total Rewards include comprehensive benefits and differentiated compensation offerings to give each employee the opportunity to be their best every day.