Staff AI Engineer, Agents and Evals

This is a fully remote position, open to applicants in United States.

📋 Description

• Take ownership of Mimo’s context layer, encompassing system context, retrieval, and memory as a unified budget.

• Establish contracts, outcomes, and failure modes for tools utilized by Mimo.

• Develop offline evaluation suites, graded live passes, and regression gates to monitor behavior changes.

• Influence how Mimo reasons, queries, explains, and recovers.

• Set rules that users can customize for their own Mimo experiences.

• Determine the deployment of models and create release gates for new model implementations.

• Collaborate with the CTO and engineers in the development of Mimo’s tools.

• Assess changes based on the tangible outputs generated by Mimo.

⛳️ Requirements

• Proven experience in building agent systems in a production environment, including tool usage, multi-step loops, streaming, and failure recovery.

• Background in designing evaluation suites, regression gates, and human review workflows.

• Familiarity with context engineering, including managing prompts, retrieval, and memory as a cohesive budget.

• Expertise in creating reliable tool contracts, schemas, outcomes, and error handling.

• Proficiency in Python.

• Experience with either a large-scale production codebase that adheres to a strict testing standard or a candidate-developed published benchmark.

• Strong technical writing skills.

• Authorization to work in the US is mandatory.

• Must not require visa sponsorship.

• US-person status may be necessary for compliance with export control regulations.

• No formal degree is required.

🏝️ Benefits

• Equity with a four-year vesting schedule and a one-year cliff.

• Comprehensive health insurance coverage.

• Dental insurance.

• Vision insurance.

• Opportunity for remote work across the United States.

• Occasional travel to Newlab in Detroit.

• Commitment to being an equal-opportunity employer.