Staff Applied AI Engineer – Healthcare Agentic Systems

Arcadia; United States; Remote; USD 175,000 - 200,000; Full time

Benefits offered by this job

Flexible, remote-friendly culture

Employee development programs

Arcadian community

Job summary

Arcadia seeks a staff-level AI software engineer to own product-layer decisions for agent behavior, including prompting, retrieval, context, memory, and escalation. You will partner with Product and Engineering to scale responsible AI in real healthcare workflows.

You will build evaluation suites, improve reliability and steerability, and ensure safety while maintaining cost-efficiency across multi-turn agentic tasks.

Qualifications

• 8+ years of production software engineering experience with 3+ years owning ML/LLM/agentic systems in regulated environments.
• Hands-on experience diagnosing why an agent failed and attributing fixes to instruction, retrieval, or memory design.
• Familiarity with RAG architecture and production-grounded evaluation frameworks.
• Experience with AWS AI/ML services (Bedrock, SageMaker) to build and evaluate in Arcadia’s environment.

Responsibilities

• Design and iterate on agent behavior across real, live workflows.
• Design retrieval, context, memory, and escalation patterns for grounded AI.
• Create prompts and templates for consistent agent behavior and scoring.
• Run evaluations measuring performance, regressions, and edge cases in production.

Skills

8+ years exp

ML/LLM expertise

Agentic systems ownership

RAG architecture

AWS Bedrock

SageMaker

evaluation frameworks

human-in-the-loop

Tools

Bedrock

SageMaker

Job description

Arcadia seeks a staff-level AI software engineer to own product-layer decisions for agent behavior, including prompting, retrieval, context, memory, and escalation. You will partner with Product and Engineering to scale responsible AI in real healthcare workflows.

You will build evaluation suites, improve reliability and steerability, and ensure safety while maintaining cost-efficiency across multi-turn agentic tasks.