Director of AI
Senior Principal AI Architect
PepsiCo · Consumer Goods / Food & Beverage · 100,000+ employees · 📍 Plano, TX
PepsiCo is building an enterprise-grade AI observability platform to govern its rapidly expanding agentic AI ecosystem — spanning LLMs, multi-agent systems, physical AI, and third-party platforms — and needs a senior architect to own that infrastructure end-to-end.
AI Maturity
advanced
Reports To
CTO
Why Role Exists
Expansion
Company Stage
public
Source
linkedin
Key Responsibilities
- Define and own enterprise observability architecture for AI agents, LLMs, multi-agent workflows, and physical AI systems
- Build unified telemetry pipelines with metrics, logs, traces, semantic signals, and real-time event streaming (Kafka) at scale
- Instrument OpenTelemetry (OTEL) across Salesforce AgentForce, ServiceNow, Microsoft Agent 365, and internal frameworks
- Design observability for agent fleets, multi-modal pipelines, self-evolving RL harnesses, and physical AI signal capture
- Lead observability-driven red team exercises targeting adversarial prompt injection, model evasion, and multi-agent trust failures
- Establish cost telemetry and FinOps observability for token consumption, inference cost allocation, and GPU efficiency
- Deliver SLO/SLA management, incident runbook automation, and RCA tooling to reduce MTTR across agentic services
Requirements
- Deep expertise in OpenTelemetry (OTEL) and distributed tracing at enterprise scale
- Hands-on experience with LLM workflows, multi-agent orchestration, and agentic AI platforms
- Proficiency with real-time streaming infrastructure (Kafka)
- Experience across major cloud platforms (Azure, AWS, GCP)
- Knowledge of AI safety, red teaming, Responsible AI (RAI), and guardrail design
- Familiarity with agent protocols (MCP, A2A, UCP, AP2)
Signals
- ⚠ No salary or compensation range disclosed
- ⚠ Posting appears truncated — responsibilities section cuts off mid-sentence
- ⚠ Onsite in Plano, TX may limit candidate pool for a highly specialized role
- ⚠ Extremely broad scope (physical AI + LLMs + RL + red teaming + FinOps) risks role becoming unmanageable for one person
- ✓ Role sits within a named, established 'AI Solutions & Platforms organization' — not a startup team
- ✓ Live integrations with Salesforce AgentForce, ServiceNow, and Microsoft Agent 365 confirm real budget and deployment scale
- ✓ Cross-domain mandate signals executive sponsorship and organizational seriousness about AI governance
- ✓ PepsiCo's scale ($91B+ revenue) means significant infrastructure investment is plausible
- ✓ Physical AI and RL harness mentions suggest differentiated, forward-looking AI investment beyond typical enterprise use cases