Director, Agentic AI Infrastructure & Orchestration Engineer

What You'll Do

• Top Three Strategic Priorities:
• 1. Architect and Deploy the Agentic Infrastructure Layer
• Design and implement a scalable runtime environment purpose-built for a media and commerce organization, including:
• Agent orchestration and task routing
• Quality gates and human-in-the-loop checkpoints
• Logging, observability, and convergence tracking
• CI/CD and deployment pipelines
• On-premise or hybrid environment architecture
• This foundation must be secure, extensible, and maintainable by the existing engineering team
 
• 2. Lead the Tool Replacement Program
• Audit current SaaS licenses and identify high-cost, high-leverage replacement opportunities (e.g., CRM, analytics, workflow tools)
• Using agentic development practices:
• Build internal tools with AI agents writing and testing code
• Maintain human review at critical checkpoints
• Deliver measurable license cost reductions
• Demonstrate infrastructure ROI within the engagement period
• The objective: prove the system’s value by using agents to build the tools themselves
 
• 3. Codify Operational Workflows into a Reusable Skills Library
• Partner across departments to:
• Identify and document high-impact operational workflows
• Encode them into version-controlled, reusable agent “skills”
• Transform tribal knowledge into durable company-owned IP
• The result: a scalable library of automated capabilities that can operate independently of the engagement
 
• Key Responsibilities:
• Architect and deploy the agentic infrastructure layer in collaboration with executive leadership and the Platform Studio engineering team
• Design a modern, growth-oriented data access layer built on existing Snowflake infrastructure and related data sources
• Enable agent-native querying, rapid experimentation, and insight generation
• Evaluate and integrate open-source orchestration tools to maintain vendor-agnostic flexibility
• Implement test-driven development enforcement, intelligent retry logic, and convergence tracking
• Establish governance frameworks with human-in-the-loop checkpoints for production deployment and auditability
• Embed seamlessly with existing engineering teams—building trust, not parallel power structures
• Develop documentation, training protocols, and knowledge-transfer systems to ensure long-term sustainability
• Partner with the COO and department leaders to prioritize automation initiatives that simultaneously reduce cost and increase velocity
• Foster a culture shift toward agentic development as a core capability within the organization

Who You Are

• Proven experience building and deploying AI agent infrastructure in production environments
• Deep proficiency across LLM orchestration frameworks, retrieval-augmented generation (RAG), prompt engineering, and agentic coding workflows
• Strong systems architecture background with infrastructure designed for long-term maintainability
• Experience with data engineering and analytics infrastructure (Snowflake and API-based systems preferred)
• Comfort evaluating and integrating open-source tooling within a multi-model ecosystem (e.g., Claude, Codex, Gemini, open-source models)
• Git-native workflow mindset (branches, PRs, CI/CD, TDD)
• Exceptional ability to translate complex architecture into business-aligned outcomes
• Experience embedding within existing teams and leaving them stronger than you found them
• Media, entertainment, or commerce industry experience preferred but not required
• Comfortable with whiteboarding systems with executive leadership
• Creating shipping production-ready pipelines
• Reviewing agent-generated code
• Ability to train teams to extend what you’ve built
 
• Success Metrics (By the end of the 12 week engagement, you will have delivered):
• A working agent runtime environment with orchestration, logging, quality gates, and CI/CD pipelines
• At least two deployed internal SaaS replacements with documented cost savings
• A versioned skills library containing a minimum of ten codified operational workflows
• A defined and partially implemented data access layer serving both human users and AI agents
• Complete knowledge transfer to the existing engineering team
• Seamless organizational integration with no structural friction
 
• What Success Looks Like:
• Success is not measured solely by architectural elegance – It is measured by operational independence
• Workflows that execute autonomously
• Tools built internally rather than rented
• Data systems that surface insights proactively
• An engineering organization fluent in agentic development