Applied AI Engineer

The University of Michigan’s Center for Academic Innovation is seeking an Applied AI Engineer for a one-year post-graduate fellowship. The role prototypes and evaluates AI-enabled workflows, supports operation of the Center’s local AI server, and shares practical guidance with staff and student fellows.

What you'll do:

• Prototype and test AI-enabled workflows for CAI projects selected through intake, building lightweight applications, connectors, retrieval pipelines, and automations that integrate with existing CAI systems
• Support pilots with domain experts, collect user feedback, diagnose failures, and iterate
• Build evaluation datasets, test cases, and repeatable harnesses for quality, reliability, safety, and cost
• Implement logging, access controls, guardrails, and operational telemetry in partnership with TIS staff
• Establish defined operating work for CAI's local AI server, including model deployment, updates, monitoring, and runbook upkeep
• Share practical guidance through demonstrations, office hours, and workshops for staff and student fellows

What they're looking for:

• A bachelor's degree in computer science, information, human-computer interaction, data science, engineering, learning sciences, or a related field, or an equivalent combination of education and applied experience, completed by the start date
• Hands-on programming experience in Python, JavaScript, or a comparable language, including experience working with APIs
• Experience building applications with large language models, retrieval-augmented generation, structured outputs, agents, or related AI techniques
• Experience using Git and working with cloud, container, or modern application-development environments
• Experience designing tests or evaluations and interpreting both quantitative results and user feedback
• Working knowledge of software security, privacy, access control, and responsible AI practices
• Experience with retrieval-augmented generation and agent frameworks, experience with evaluation or observability tooling
• Ability to learn and appropriately use new technologies to improve processes, communication, and outcomes, including AI-enabled tools where relevant to the role, consistent with university policy and job responsibilities
• Candidates must have legal authorization to work in the United States

Benefits:

• Hybrid work: minimum of 4 days in the office per week, Monday through Thursday, with an option of remote work on Fridays; may be required to work onsite on Fridays if necessitated by the center's policy and domain leadership or by job requirements.