Identifying high-quality educators at the point of hire can reduce future recruitment costs and minimize the impact of attrition on school organizations and student learning. One low-cost way to screen applicants and learn about their beliefs, values, and pedagogy is through their short-essay writing samples. However, there is limited research linking short-essay data with educators' hiring and on-the-job outcomes. The present study analyzes six years of educator applicant data from a large urban school district in California, aiming to understand whether applicant essays are associated with their search, persistence, and performance outcomes. Essay quality and theme indicators identified by machine learning techniques significantly predict essay scores, which in turn significantly predict teacher hiring status and on-the-job outcomes. Results are comparable in magnitude to several more expensive screening procedures. We also find smaller, suggestive relationships between certain essay themes and educator hiring status and value-added, but not retention. While the themes and qualities that are most advantageous for hiring and job performance may vary by school district, our findings suggest that using essays aligned to a district's specific priorities has the potential to enhance teacher recruitment, retention and improve student achievement.