Circana is building an agentic ecosystem on the industry's richest view of consumer behavior: intelligent systems that reason, plan, and act across the data the world's largest brands and retailers run on. AI\\Agentic Architects lead that work solution by solution. Embedded with the experts who hold the domain knowledge, they design the agents, set the patterns, and own the outcome from first conversation to production use. Each architect anchors a pod of engineers and carries the technical direction for a solution family for the length of a build.
Responsibilities
Lead the design and delivery of agentic systems for a solution family, from domain discovery through production use.
Translate domain expertise into intent ontologies, tool contracts, and evaluation sets, working directly with subject matter experts.
Own harness engineering: context management, tool contracts and MCP servers, guardrails, retries and graceful degradation, tracing and replay.
Establish evaluation as infrastructure: golden traces, regression suites, routing tests, offline and online measurement.
Set reusable patterns (shared services, tools, and conventions) that make every subsequent build faster than the last.
Anchor a pod of engineers: set technical direction, review the work, and raise the bar.
Represent the work credibly to product owners, architects, security, executives, and client stakeholders.
Requirements
A strong track record of designing and shipping agentic systems into production for real users.
Deep production fluency with agent orchestration frameworks: LangGraph, Google ADK, or Microsoft Agent Framework.
Production experience across multiple model families, with clear judgment on when to reach for which.
Production experience on at least one major cloud AI stack.
Mastery of the practical craft: context management, prompt design, plan generation, structured outputs, and tool design.
Daily fluency with agentic coding tools; Claude Code preferred.
Command of the agentic development lifecycle: specs as source of truth, evals as continuous integration, prompt and tool versioning, and trace-driven debugging.
Communication that lands at every altitude: engineers, domain experts, and senior executives alike.
Nice-to-haves
Consumer goods, retail, or large-scale measurement data background.
Prior forward deployed or field engineering experience at an AI or data company.
Public artifacts: talks, open source, technical writing.
