Resumes

Finding Good Resumes

Sector: Recruitment

Capability: Document Labeling

When it comes to constructing a resume, it's often unclear which parts are helping your chances of getting an interview and which parts are hurting them. For example, Bella has been searching for Data Analyst jobs in the Seattle area. She has manually applied to around 1,500 job opportunities but hasn't had any luck finding a full-time position. In Bella's own words, "The job-seeking process is literally EXHAUSTING. I am just mentally drained but cannot give up." Bella seeks advice on how to create a better resume that will catch the attention of recruiters.

Training an AI algorithm to predict good resumes can be challenging and prone to bias. For example, if the 50 labeled good resumes happen to be from individuals in Indiana, the model might mistakenly assume that candidates from Indiana generally have good resumes. In many cases, there are only a few labeled data points representing candidates who made it to interviews, while a large amount of data remains unclassified, ignored, and unlabeled.