Principal Agentic Systems Architect

About the Role: Grade Level (for internal use): 12The Role: Defines and guides the architecture of enterprise agentic systems and AI-enabled engineering capabilities. Designs reusable architectures and engineering patterns so teams can build, integrate, evaluate, secure, and operationalize agents at scale, independent of the specific framework, model, or runtime chosen. Hands-on, working across multiple initiatives to turn business and technical requirements into production-ready solutions.  Key Responsibilities  Define reference architectures and reusable design patterns for agent development, orchestration, tool calling, context management, multi-agent interaction, and human-in-the-loop controls.  Set integration patterns for coding agents, agent frameworks, MCP servers, enterprise APIs, and security tools, so new agents onboard without custom one-off plumbing.  Architect isolated, ephemeral execution environments with identity and access controls, secrets management, and full auditability.  Establish engineering standards for the agent lifecycle: CI/CD, versioning, behavioral testing, agent evaluations, release management, and runtime operations.  Define approaches for observability, failure handling, and production readiness across agent workloads.  Guide engineering teams, review solution designs, evaluate technology choices, and resolve complex architectural challenges across initiatives.  Validate architectural decisions through prototypes and proofs of concept alongside platform engineers and delivery teams.  Document patterns and standards so they are reused across projects and teams, not rebuilt each time.  Core Skills Required  Hands-on experience designing and building LLM-based applications, AI agents, or agentic workflows.  Strong understanding of agent frameworks, coding agents, tool integration, MCP, and orchestration patterns.  Cloud-native architecture experience: APIs, containers, Kubernetes, and distributed systems.  Application security: IAM, secrets management, secure SDLC, and software supply-chain security.  CI/CD, automated testing, observability, and production engineering practices.  Strong programming skills in Python, Java, or Go, sufficient to build working prototypes.  Architectural decision-making and clear technical communication, with the ability to influence multiple teams without direct authority.  Flexibility to work across varied and evolving tech stacks, not tied to one runtime, framework, or model provider.  Qualifications  10+ years in software engineering, solution architecture, or distributed systems, including 2+ years building LLM or agent-based systems.  Experience in secure/regulated enterprise environments preferred.  Nice to have: experience with enterprise coding-agent platforms (e.g., Devin, GitHub Copilot agents, Claude Code) or similar AI-assisted engineering tools.  Nice to have: experience building secure, ephemeral execution environments for autonomous or semi-autonomous agents.