building a evaluator transcript and figuring out how to use BrightData tools
pass in an interview transcript and it outputs a rubric with a second agent evaluating the first agent on how the first agent is doing