We are looking for a Senior Agentic AI Engineer to design, build, and operate multi-agent LLM systems that automate complex, document-heavy business workflows.
You will own agentic applications end to end, from agent design and tool integration through evaluation, deployment, and ongoing operation on AWS.
Your first focus will be taking ownership of an existing multi-agent platform and scaling it, then helping shape the next wave of agentic solutions across the organization.
What you will do
- Design and build multi-agent systems: orchestration, tool use, memory, and human-in-the-loop approval steps.
- Build RAG and document-intelligence pipelines that turn unstructured documents (PDFs, tables, scanned files) into reliable structured data.
- Create evaluation frameworks and regression tests that measure agent accuracy, recall, and groundedness, and catch silent failures.
- Run agentic systems in the cloud: deployment, observability, incident response, and cost and latency optimization.
- Make architecture decisions on model selection, prompt and context design, guardrails, and safety controls.
- Partner with domain experts, product, and platform teams to turn business processes into agentic solutions.
- Set engineering standards for agentic AI and mentor other engineers.
What you bring
- 13+ years of software engineering experience, with strong 5+ years of experience with Python and testing practices.
- 2+ years building LLM applications deployed to real users, including agents, tool calling, and RAG.
- Hands-on experience with at least one agent framework (LangGraph, Strands, CrewAI, AutoGen, OpenAI Agents SDK, or similar) and with tool protocols such as MCP.
- Experience evaluating and improving LLM output quality in a systematic way.
- Solid AWS experience (Bedrock, Lambda, Step Functions, DynamoDB, S3, IAM) and infrastructure as code (Terraform preferred).
- AWS Bedrock AgentCore, Bedrock Knowledge Bases, or vector search (OpenSearch, pgvector).
- Clear communication with both engineers and non-technical stakeholders.
Nice to have
- Document AI: Textract, OCR, layout and table extraction, vision-language models.
- Experience with Anthropic Claude models, prompt caching, and model tiering for cost.
- Background in regulated domains such as healthcare, insurance, or claims.
- CI/CD for AI systems (Jenkins, GitHub Actions) and basic React.
Rate:
Not specified
Location:
Remote
IR35 Status:
Outside
Remote Status:
Remote
Industry:
AI & Machine Learning
Seniority Level:
Senior