Senior AI Engineer, Architect

PepsiCo•Plano, TX

About The Position

We are seeking a Senior AI Engineer to define and drive the end-to-end engineering of an enterprise-grade agentic orchestration capability that enables smart AI agents to autonomously execute workflows, collaborate with humans, and operate securely with governed access. This role owns the technical direction and delivery of core capabilities spanning agent workflow development environments, automated CI/CD and safe migration patterns, human–agent collaboration and long-running orchestration, and agent identity/registry/marketplace with policy enforcement. You will serve as the technical authority—establishing standards for reliability, auditability, security, and performance; driving cross-team execution; and ensuring adoption at scale through enablement and strong operational practices.

Requirements

Bachelor’s in CS/AI/ML/Data Science or equivalent experience required.
10 year experience in ML, Data Science, AI required.
Extensive experience designing and operating enterprise platforms/services with production reliability and governance requirements.
Systems/platform architecture: multi-tenant isolation, scalability, versioning, backward compatibility, release sequencing
Orchestration and workflow systems: Temporal-class systems (or equivalent) including long-running workflows, compensation, state persistence
Identity and security architecture: SSO (SAML/OIDC), non-human identity, RBAC/ABAC, consent propagation, secrets/keys rotation, least-privilege design
Governance and compliance engineering: audit logging models, approval workflows, policy routing, PII redaction, retention/purge controls
Observability/SRE partnership: SLO definition, OTel-based telemetry, incident management, reliability engineering
Developer enablement: SDK design, reference implementations, platform adoption strategy, mentoring and technical leadership

Nice To Haves

Master’s preferred

Responsibilities

Define reference architecture, design standards, and engineering guardrails for agent workflow orchestration, human collaboration, and identity/governance capabilities.
Own sequencing of releases, deprecation strategy, and compatibility standards to enable safe evolution with minimal disruption.
Establish and enforce non-human identity patterns, consent propagation mechanisms, RBAC/ABAC policy models, and least-privilege access across agent workflows.
Ensure end-to-end auditability for agent actions, prompt/tool changes, model switches, handoffs/messages, approvals, and access decisions; define evidence requirements for compliance.
Define and enforce data classification, PII redaction, retention/purge, and policy-based routing to compliant models/providers.
Define and drive implementation of deterministic handoff patterns (assign/escalate/co-pilot/co-author), resilient messaging, and stateful long-running workflows with timers and compensation/rollback.
Ensure seamless integration into enterprise systems (CRM/ITSM/custom apps) via gateways and standardized interfaces.
Define promotion gates and automated CI/CD standards including versioning, testing, security scans, approvals, and drift detection.
Drive safe migration practices between model providers/versions with minimal downtime and proven rollback; define operational playbooks.
Own SLIs/SLOs and operational posture: observability standards (metrics/logs/traces), incident and credential compromise runbooks, and release readiness reviews.
Deliver enablement: reference implementations, developer playbooks, training for platform ops and application teams; mentor senior and junior engineers.