Staff Software Engineer, Agentic Applications

Databricks Databricks · Data AI · Mountain View, CA · Engineering

Staff Software Engineer role focused on building and owning LLM-powered agentic workflows for marketing content creation and publishing. The role involves defining agent architecture, ensuring reliability, implementing evaluation frameworks, and encoding guardrails within the marketing content lifecycle.

What you'd actually do

  1. Own the architecture and delivery of agentic workflows that enable marketing to self-serve content creation and publishing across blogs, landing pages, microsites, and social, working directly with marketing stakeholders to understand their workflows, define quality criteria, and build tools they trust and use
  2. Build and own the agent reliability system for this domain: define output quality criteria per workflow, implement eval suites and behavioral regression testing, and build feedback loops that improve agent output over time based on what marketers approve, edit, and reject
  3. Architect human in the loop review as a system property across the full publishing pipeline: proposal and commit separation, diff-based review UI, approval routing, notification system, and audit logging that govern when agents publish autonomously and when human judgment is required
  4. Encode brand, legal, and compliance constraints as enforceable guardrails within agent workflows, ensuring agents operate within defined boundaries without requiring human intervention on every decision
  5. Mentor and grow the Senior Engineer on the team, setting the standard for production-grade agentic engineering at Databricks

Skills

Required

  • LLM-powered systems in production
  • human in the loop system design
  • agent interaction with content systems
  • web infrastructure including CMS, deployment pipelines, and edge delivery
  • mentoring engineers

Nice to have

  • Translate workflow requirements into agent architecture
  • work directly with non-engineering stakeholders

What the JD emphasized

  • 12+ years of software engineering experience with a clear track record of technical ownership on production systems
  • Hands-on experience building and operating LLM-powered systems in production
  • strong instinct for designing quality and evaluation frameworks for probabilistic or generative outputs
  • Deep understanding of how agents interact with content systems, publishing pipelines, and web infrastructure including CMS, deployment pipelines, and edge delivery

Other signals

  • LLM-powered agents
  • agentic SDLC
  • orchestrated, continuously improving system
  • agent reliability system
  • human in the loop review