Senior AI Applications Engineer, Agentic LLMs

Booz Allen Hamilton; Washington; On-site; USD 87,000 - 198,000; Full time

Job summary

Booz Allen Hamilton is seeking a Senior Software Development Engineer to design and implement agentic AI workflows on GRACE, ARPA-H's production AI assistant. You will own the application-layer orchestration, tool-calling, prompts, and multi-agent coordination, collaborating across product, UX, and operations to ship reliable features.

You’ll build end-to-end capabilities from backend logic to API contracts consumed by the frontend, ensuring security, observability, and cost-efficient prompts.

Qualifications

• 7+ years of software engineering experience
• Experience with Python and another backend language
• Experience with cloud platforms (AWS, GCP, Azure)
• Experience with containers and CI/CD pipelines
• Knowledge of distributed systems, APIs, data pipelines and design patterns
• Ability to communicate clearly and work with urgency

Responsibilities

• Design and build GRACE’s core agentic workflows
• Implement and evolve A2A communication patterns at the application layer
• Build and maintain the tool-calling layer with definitions, schemas, error handling
• Own MCP client-side integration and tool discovery/invocation
• Design multi-agent workflows that are observable and debuggable in production
• Own LLM orchestration at the application layer and prompt management
• Build and maintain RAG features and grounding/citation strategies
• Prototype agentic features and iterate based on user data

Skills

Python

Backend development

Cloud platforms

Containerization & CI/CD

Distributed systems

APIs & data pipelines

Async programming

Communication

Self-starter

Education

Bachelor's degree in Computer Science or Software Engineering

Tools

MCP

Job description

Booz Allen Hamilton is seeking a Senior Software Development Engineer to design and implement agentic AI workflows on GRACE, ARPA-H's production AI assistant. You will own the application-layer orchestration, tool-calling, prompts, and multi-agent coordination, collaborating across product, UX, and operations to ship reliable features.

You’ll build end-to-end capabilities from backend logic to API contracts consumed by the frontend, ensuring security, observability, and cost-efficient prompts.