Log In
Sign Up
Company Description
ReasonCore AI is a dynamic and innovative organization focused on advancing the field of artificial general intelligence. We pair proprietary data-engine technology with a global network of vetted data experts to produce the high-end reasoning content that internet-scale data simply cannot supply — across math, scientific, enginering, financial, coding, and spatial domains. Headquartered in the San Francisco Bay Area, ReasonCore AI is committed to fostering creativity, collaboration, and adaptability. Joining our team means contributing to impactful projects that push the boundaries of AI innovation.
The role
You will be one of the experts on the front line of our data engine: designing, authoring and reviewing reasoning content that frontier AI labs use to train their best models. This is not labeling. This is advanced LLM reasoning work — constructing multi-step problems, producing reference solutions, and helping us define what “a correct answer” looks like in your domain.
What you will do
· Author original reasoning problems and reference solutions, calibrated to lab-defined difficulty bands.
· Review and rate model responses on rubric criteria; identify failure modes and edge cases.
· Contribute to private benchmarks and Eval-as-a-Service datasets shipped to frontier labs.
· Document your work clearly enough that it can train and evaluate models for years.
Whom are we looking for
· Math (or closely related quantitative field such as Economics, Engineering) Bachelors, Master’s, or PhD graduates (or currently enrolled in an accredited university).
· Strong English written communication; clear, unambiguous, structured.
· Comfort working independently against deadlines in a fully remote environment.
· Curiosity about how frontier AI models actually learn — you don’t need ML experience, but you should be excited to develop intuitions on how LLMs work.
Nice to have
· Prior teaching, tutoring, or exam-writing experience.
· Some exposure to LLMs (you’ve prompted models for your work, even informally).
· Multiple-domain expertise (e.g., a physics expert who also codes well).
What we offer
· Fully remote, flexible hours (we coordinate across time zones via async-first workflows).
· Competitive compensation depending on your qualifications.
