There are several ways to assess AI fluency on our platform and it depends on what the role is and what you need graded and analyzed. More information on each below.
Technical and coding roles
For technical roles like software engineering, data science, machine learning, front-end developer, etc. the best way to measure AI fluency is to have the candidate solve a take-home project with AI assistance enabled. You can read more about our AI assistant that comes with ChatGPT and Claude built-in here. You can create custom projects in our native IDE or in VS Code with comes with Codex, Claude Code, and our own AI assistant.
We offer dozens of take-home projects that are built specifically to be solved with AI. The final grade and analysis looks at how the candidate wrote prompts, what they accepted or rejected from the agent, and how they explain their final work in a README file.
Additionally, you can also create free-form questions where the candidate has to explain their work or parts of their code, and you can have AI grade this and provide a score.
Non-technical roles
For non-technical roles, we also provide an environment where candidates can work with AI tools such as Codex/ChatGPT or Claude Code to put together documents, slideshows, and even upload prototype apps that they can share internally.
Below are some examples for marketing and product analytics roles where the candidate was tasked with follow instructions from an internal document. The project templates each consisted of a document with requirements, and the candidates had to work with AI to analyze data, generate new charts and documents.
Task to create an internal app prototype for marketers
Task to analyze a CSV of survey responses and generate a slideshow
