Principal Software Engineer - Agentic Validation, Trust & Evaluation

Employee Compute is evolving from endpoint operations to an AI-enabled compute control plane that supports a global workforce at scale.

As Executive Director for Agentic Validation, Trust & Evaluation , you will own and lead the function that verifies AI-enabled employee experiences are safe, explainable, policy-compliant, and operationally ready before and during production use.

This role sits within Risk & Controls and is accountable for setting the strategy for, building, and running the validation, trust, and evaluation operating model for agentic capabilities used across Employee Compute journeys.

You will partner across engineering, cybersecurity, architecture, and control teams to establish measurable confidence in AI-assisted workflows through strong evaluation frameworks, auditable evidence, and clear governance outcomes aligned to a highly regulated environment.

This role holds executive-level individual accountability for the effectiveness of the agentic validation and control environment within Employee Compute, including timely identification, escalation, and remediation of control gaps.

The Executive Director is the senior named control owner for go/no-go release decisions on AI-enabled employee experiences and is answerable to senior leadership, audit, and regulators for the integrity of the evidence supporting those decisions.

Job Responsibilities



* Set and own the enterprise strategy for Agentic Validation, Trust & Evaluation for Employee Compute, translating policy and risk expectations into practical validation and control mechanisms for AI-enabled employee workflows.



* Design and operationalize end-to-end evaluation frameworks for agentic capabilities, including pre-release validation, post-release monitoring, drift detection, and periodic control attestations.



* Build standardized trust criteria for AI-assisted outcomes such as identity integrity, authorization correctness, data-handling boundaries, provenance, explainability, and reproducibility of decisions and actions.



* Establish and govern quality gates and go/no-go decision practices for AI-enabled releases, ensuring evidence-based readiness across security, controls, resiliency, and operational support.



* Own the mandate that validation and control requirements are embedded in design and delivery lifecycles, with authority to block releases that fail to meet engineered-in control standards (controls engineered in, not retrofit after launch).



* Create and own KPI/OKR and scorecard structures for trust and evaluation outcomes, including control health, issue closure velocity, incident learnings, and model/tool performance against policy expectations.



* Own and personally attest to examiner-ready evidence packages and control-testing results presented at risk and control forums, including executive communication on material risks, decisions, and mitigations to senior technology and control leadership.



* Be accountable ...