AI Engineer (Agentic AI / AWS Bedrock)

TheStaffed•Los Angeles, CA

2d•Hybrid

About The Position

Our client is seeking an AI Engineer to design, develop, and deploy intelligent AI agents using Claude Agent SDK, Python, and Amazon Bedrock. The role involves building production-ready agentic systems, autonomous workflows, and integrations with enterprise applications, data sources, external tools, APIs, knowledge bases, and foundation models. The engineer will build agent orchestration workflows on Amazon Bedrock AgentCore Runtime, with error handling, retries, and guardrails. The position also covers LLM and agent observability, production evaluation frameworks, scalable RAG pipelines, intelligent document processing, CI/CD automation, and AWS security and cost optimization.

Requirements

6–10 years of professional software engineering or AI/ML experience.
Demonstrated expertise in Python and LLM-based systems.
Advanced proficiency with Claude Agent SDK, agent orchestration, tool calling, and agentic AI frameworks such as LangChain and LangGraph.
Proven experience deploying AI agents in production environments.
Deep expertise with Amazon Bedrock, Bedrock AgentCore Runtime, and AWS services including Step Functions, Textract, Comprehend, OpenSearch, IAM, and CloudWatch.
Hands-on experience with LLM observability and evaluation tools, including Langfuse, LLM-as-a-Judge frameworks, RAGAS, DeepEval, and custom evaluation metrics.
Strong proficiency with vector databases and RAG implementation, including Pinecone, OpenSearch, FAISS, or pgvector; experience with embeddings, semantic search, and retrieval optimization.
Expertise in CI/CD and DevOps, including GitHub Actions, Git, Docker, Terraform, Infrastructure as Code, and automated deployment pipelines.
Proficiency in prompt engineering, Model Context Protocol (MCP), AI governance principles, and AWS architecture best practices for secure, scalable systems.
Must be available for a hybrid work arrangement.
W2 employment only.

Responsibilities

Design, develop, and deploy AI agents using Claude Agent SDK and Python.
Build autonomous workflows that execute complex tasks and interact with external tools, APIs, and knowledge bases.
Develop agent orchestration workflows on Amazon Bedrock AgentCore Runtime, including error handling, retry mechanisms, and guardrails.
Integrate agents with enterprise applications, data sources, and foundation models.
Implement observability and monitoring using Langfuse to trace prompts, responses, tool calls, execution metrics, and production performance.
Design and implement continuous evaluation frameworks for production AI agent traffic using LLM-as-a-Judge, RAGAS, DeepEval, and custom metrics.
Build scalable RAG pipelines and vector database solutions for document ingestion, chunking, embedding generation, and knowledge base retrieval optimization.
Develop AWS Step Functions workflows integrating Amazon Textract and Amazon Comprehend for intelligent document processing and data transformation.
Build and maintain GitHub Actions CI/CD pipelines for agent code builds, testing, deployment, and integrated evaluation workflows for Amazon Bedrock AgentCore Runtime.
Implement AWS security best practices, including IAM policies, access controls, authentication, CloudWatch monitoring, and cost optimization for production agent systems.