Frontier AI Evaluation Engineer - MTS (Remote)

micro1 Inc.; United States; Remote; USD 240,000 - 350,000; Full time

Benefits offered by this job

Equity compensation

Bonus opportunities

Health insurance premium reimbursement

401(k) with company match

Paid time off

Remote-first workforce

Job summary

micro1 is seeking a Member of Technical Staff for Realm Lab to drive frontier AI research initiatives from a remote setting in the United States. This hands-on role sits at the intersection of ML research, evaluation engineering, and data science to build benchmarks, RL environments, and high‑quality datasets that advance model capabilities.

You will collaborate with researchers and domain experts to design experiments, validate evaluations, and develop reproducible evaluation pipelines,

Qualifications

• Strong background in computer science, ML, or AI.
• Hands-on experience designing ML evaluations and benchmarks.
• Proficient in Python and building evaluation tools.

Responsibilities

• Own research and evaluation initiatives end-to-end.
• Design and validate capability evaluations and benchmark suites.
• Develop datasets, task specs, evaluation rubrics.
• Analyze model behavior to identify capabilities gaps.
• Collaborate with researchers to design experiments.
• Build reproducible evaluation pipelines and feedback loops.
• Contribute to benchmarks, papers, and open artifacts.

Skills

Machine Learning

Evaluation research

Python

Data pipelines

LLMs understanding

Experiment design

Education

BS/MS in CS/ML

Tools

Python

Benchmark suites

RL environments

Data pipelines

Job description

micro1 is seeking a Member of Technical Staff for Realm Lab to drive frontier AI research initiatives from a remote setting in the United States. This hands-on role sits at the intersection of ML research, evaluation engineering, and data science to build benchmarks, RL environments, and high‑quality datasets that advance model capabilities.

You will collaborate with researchers and domain experts to design experiments, validate evaluations, and develop reproducible evaluation pipelines,