Enterprise AI Orchestration Architect, Senior Manager

Job Number: R0251883

About the role

In this role, you’ll build reusable agents and intelligent applications that perform enterprise work reliably. You will develop agent behavior, orchestration, tool integrations, and evaluation capabilities that support customer solutions from demonstration through production. You will translate workflows into tested software using model APIs, retrieval, tool calling, memory, and human oversight. You will instrument execution, maintain evaluation datasets, and improve quality through trace analysis and regression testing. Partnering with Platform, Data, and Forward-Deployed Engineering, you will deliver reusable components, integration contracts, and deployment guidance. You will own agent behavior and application quality, including the evidence needed to validate them.

Responsibilities

• Build reusable agents and intelligent applications for enterprise workflows
• Develop agent behavior, orchestration logic, tool integrations, and evaluation capabilities
• Translate workflows into tested software using model APIs, retrieval, tool calling, memory, and human oversight
• Instrument execution, maintain evaluation datasets, and improve quality through trace analysis and regression testing
• Partner with Platform, Data, and Forward-Deployed Engineering to deliver reusable components, integration contracts, and deployment guidance
• Own agent behavior and application quality, including validation evidence

Requirements

• 10+ years of experience building maintainable applications, APIs, and automated tests using version control and software engineering practices, including proficiency in Python or TypeScript
• Experience integrating large language models into applications, including prompt and context management, structured outputs, and tool calling
• Experience implementing agent orchestration, workflow state, retrieval-augmented generation, and enterprise API integrations
• Experience instrumenting model calls, tool execution, state transitions, and outcomes with logs or distributed traces
• Experience developing evaluation datasets and regression tests for task success, groundedness, safety, tool correctness, latency, and cost
• Ability to design appropriate human approval steps, tool permissions, input and output validation, error handling, and fallback behavior
• Ability to diagnose application failures and distinguish agent behavior defects from data, platform, or customer configuration issues
• Ability to document interfaces and limitations and collaborate on customer readiness and reusable engineering improvements
• Bachelor’s degree

Preferred Qualifications

• Experience with LangGraph, Semantic Kernel, MCP, OpenTelemetry, LangSmith, Langfuse, or comparable technologies
• Experience with multi-agent coordination, memory design, reusable SDKs, or reference architectures
• Experience delivering full-stack AI applications or improving production workflows through evaluation and trace analysis

Benefits

• Health, life, disability, financial, and retirement benefits
• Paid leave
• Professional development and tuition assistance
• Work-life programs and dependent care
• Recognition awards program for exceptional performance and demonstration of values

Pay

The projected compensation range for this position is $162,800.00 to $303,000.00 (annualized USD). Salary is determined by factors including location, education, knowledge, skills, competencies, experience, contract-specific affordability, and organizational requirements.

Work Model

Remote: If this position is listed as remote, there may still be occasions when you are required to work in person at a Booz Allen or customer facility. Hybrid: If this position is listed as hybrid, you will be expected to work from a Booz Allen or customer facility, in alignment with your leadership's expectations and the needs of the role. Onsite: If this position is listed as onsite, work will primarily be performed full-time at a customer facility, where employees will collaborate directly with colleagues and customers as required by the role. Employees working virtually are generally expected to have their cameras on during meetings.