QUALIFICATIONS

This technology position is designed for candidates in the humanities. It focuses more on your skills and abilities than on your formal experience, offering a rare opportunity that suits sophomores, recent graduates, career changers, and anyone looking for an engaging remote role.

You naturally have a strong ability to write compelling copy with an intuitive, creative flair.

You also have the metacognitive ability to “show your work.” In addition, you’re highly interested in linguistics.

RESPONSIBILITIES:

As a Remote AI Content Evaluator, you’ll work with a team of other trainers who follow protocols created by leading AI researchers. You’ll help train the AI to read, write, summarize knowledge, and interpret meaning.

Your role will be to train, evaluate, and test the AI’s conversational skills—continually preparing it to meet its goals.

Most of your time will be spent creating examples of ideal conversations. You’ll act as both the User and the AI so the system can learn from your responses. You’ll also collect sources to help the AI interpret large parts of humanity’s documented knowledge, including identifying which statements are presented as facts, which are provided as context, and which recurring patterns reflect typical behavior.

For example, you’ll evaluate not only whether the AI’s factual answers are correct, but also whether it understands and interprets those facts properly.

As the project advances, you’ll also try to “break” the AI by intentionally prompting it to make mistakes, and then improving how the models recover.

This is essential to its safety, accuracy, and usefulness in real-world applications.

You’ll document failures and will have the chance to recommend improvements to our training methods to both our team and our client.

Earnings:
Hourly rate: US$22–32/hour, depending on your experience