At Workiva, we build and scale an enterprise platform that powers complex financial reporting, compliance, and collaboration for some of the world’s largest organizations. We solve for high-throughput data, strict consistency, and architectural scale while exploring what AI can do to transform how engineering teams work.
We are not simply adding AI to existing workflows. We are redesigning how engineering gets done. As agents take on more of the work, the human review steps that once caught problems are going away. AI failures often happen when nothing breaks: the model does exactly what it was asked, and the result is still wrong. That makes safety a control problem, not a prediction problem.
We are hiring a Principal Software Engineer to lead quality and safety for AI-native delivery. Our standard is quality through systems, not heroics, so the safe path is the easy path and higher velocity never means higher production risk. This is a hands-on role for someone who can define strategy, propose original solutions, and raise the technical bar for the teams around them.
What You’ll Do
Deliver outcomes
Partner with engineering, platform, InfoSec, SRE, and Compliance teams to build layered safety controls, so systems stay safe even when multiple safeguards fail.
Build pre-production verification, including evals, invariant tests, and risk-based test selection under a non-skippable invariant floor.
Build automated safety valves into development and delivery, so safety does not depend on slow human review.
Define and drive technical strategy
Lead the “Ratify Continuously, Expose Deliberately” model: continuous shadow soak, ratified releases, and deliberate per-tenant exposure.
Define the invariants our systems must never violate, and design controls that make those states unreachable.
Build Agentic Systems to remove CI bottlenecks by quarantining flaky tests and moving full suites to nightly runs.
Frame ambiguous safety problems and propose original approaches to solve them.
Increase agent autonomy only as verification improves.
Risk based evaluation approval of PRs
Design AI-driven systems
Enforce integrity-protected channels, where authorization comes only from signed sources and permission brokers, and all agent inputs and tool outputs are treated as data.
Ensure agents have zero authority over the guardrails that constrain them, such as CI gates, branch protection, and audit loggers.
Require privileged actuators to independently verify an agent’s claims about environment, approvals, and test status before acting.
Test, monitor, and observe AI models in real environments, with containment and rollback when behavior is unsafe.
Influence and lead
Raise the technical bar by making quality the default rather than the result of heroics.
Mentor and challenge Staff+ engineers to expand their scope, judgment, and impact.
Partner with engineering and product leadership to balance velocity with production risk.
Create durable artifacts such as runbooks, decision logs, and tests that can be handed off and evolved by others.
Contribute across product domains, bringing context from one area to solve problems in another.
What You’ll Need
Minimum Qualifications
8+ years of experience building and scaling complex software systems.
Bachelor’s degree in Computer Science, Information Security, or a related field; or 6+ years and a Master’s degree; or a PhD with 4+ years of experience; or equivalent practical experience
Demonstrated experience using AI as a primary development collaborator.
Preferred Qualifications
Experience designing safety-critical, high-assurance, or redundant systems.
Experience building CI/CD and test infrastructure that improves engineering velocity and quality at scale.
Experience designing evals for LLMs or AI agents.
Experience with observability and monitoring for production systems.
Experience with security controls such as least-privilege access, signed provenance, and policy-as-code.
Experience with progressive…