AI Engineer - Multi-Agent LLM Systems for Government

Uncover; Pittsburgh, Northern (Allegheny County, KY); Hybrid; USD 120,000 - 190,000; Full time

Job summary

Air seeks an experienced AI Engineer to join our Agentic AI team in Pittsburgh, PA, scaling Ace across US government use and coordinating with human users as needed. The role emphasizes planning, reliable long-horizon execution, memory, and inter-agent coordination.

Ideal candidates will showcase a project with LLMs, experience with agent frameworks (Claude/OpenAI), observability, telemetry, and rigorous evaluation. Travel up to 25% may be required.

Qualifications

• U.S. Citizenship is required.
• BS/MS/PhD in CS/CE/Data Science or related field.
• 1+ year of ML/LLM production experience.
• Hands-on with ML systems deployment and productionizing.
• Proficient in Python.
• Experience with LLM algorithms and cloud infra (AWS).
• Git proficiency and collaboration skills.
• Desire to learn new tech and work independently.

Responsibilities

• Designs and builds multi-agent systems with tool use, memory, routing, and planning.
• Develops agent skills and tools as modular services interacting with models and data.
• Owns architecture and engineering standards for agentic systems for scalability and reliability.
• Assists automated evaluation of agents, skills, and prompts across the lifecycle.
• Collaborates with Product/Implementation to apply agents to customer problems.

Skills

ML/LLM production

Python

AWS

Git

Autonomy

Team collaboration

Cloud infrastructure

Learning mindset

LLM algorithms

Teleology/Telemetry

Education

Bachelor's/Master’s/PhD in CS/CE/Data Science

Tools

Claude/GenAI SDK

Job description

Air seeks an experienced AI Engineer to join our Agentic AI team in Pittsburgh, PA, scaling Ace across US government use and coordinating with human users as needed. The role emphasizes planning, reliable long-horizon execution, memory, and inter-agent coordination.

Ideal candidates will showcase a project with LLMs, experience with agent frameworks (Claude/OpenAI), observability, telemetry, and rigorous evaluation. Travel up to 25% may be required.