Autonomous AI Agent Engineer

Booz Allen Hamilton; Alexandria (VA); On-site; USD 78,000 - 176,000; Full time

Benefits offered by this job

Health benefits

Paid time off

Job summary

Booz Allen Hamilton is seeking an Agentic AI Engineer to help design autonomous, goal-oriented AI systems for mission-ready applications. You will implement architectures that reason, plan, and act, using LangChain, LangGraph, and AutoGen, while enabling multi-agent collaboration across MCP and A2A protocols.

Role involves RAG pipelines with knowledge graphs, edge-optimized SLM tuning, and rigorous evaluation for production readiness.

Qualifications

• 2+ years of software development experience.
• 2+ years as ML engineer building production ML solutions with Docker or Kubernetes.
• 2+ years with LangChain, LangGraph, AutoGen, PydanticAI, CrewAI, or LlamaIndex.
• 2+ years cloud experience (AWS/Azure) for scalable apps.
• Experience with MCP/A2A, RAG, KG (Neo4j/NebulaGraph).
• Fine-tune LLMs/SLMs with Hugging Face, PEFT, LoRA; observability with Grafana/Langfuse/LangSmith/Phoenix.

Responsibilities

• Design and implement intelligent agent architectures that can reason, plan, and act.
• Develop multi-agent systems using MCP and A2A protocols.
• Build RAG pipelines with Knowledge Graphs for context retention.
• Fine-tune SLMs for edge deployment (ONNX/GGML/Ollama).
• Create evaluation frameworks for reliability, safety, and production readiness.

Skills

Software development

ML engineering

LangChain

LangGraph

AutoGen

A2A protocols

RAG pipelines

Edge deployment

Docker

Kubernetes

Education

Bachelor’s degree

Master’s degree

Doctorate degree

Tools

Docker

Kubernetes

Neo4j

NebulaGraph

llama.cpp

vLLM

LangChain

LangGraph

AutoGen

Job description

Booz Allen Hamilton is seeking an Agentic AI Engineer to help design autonomous, goal-oriented AI systems for mission-ready applications. You will implement architectures that reason, plan, and act, using LangChain, LangGraph, and AutoGen, while enabling multi-agent collaboration across MCP and A2A protocols.

Role involves RAG pipelines with knowledge graphs, edge-optimized SLM tuning, and rigorous evaluation for production readiness.