Capital Technology Group is a rapidly expanding technology firm and we are constantly trying to improve our own internal processes. We have been inundated with numerous software engineer resumes, which has taxed our internal HR team's bandwidth and caused substantial delays in candidate evaluation. The challenge was to reduce the time and resources spent on initial resume screening without compromising the accuracy of candidate qualification.
Digital Transformation, Software Development
Python, Data Preprocessing, Privacy Protection, Large Language Models, Prompt Engineering, Model Evaluation, API Integration
CTG initiated a two-pronged approach: 1) Utilize Python to preprocess existing resumes by stripping out personal identifiers to maintain applicant privacy, and 2) Employ prompt engineering and large language model evaluation to mimic the qualification assessments made by our HR team. Various prompts and models were tested, with the most effective combination selected for production. This new system has been integrated into the existing application/HR software, automating the daily processing of new applications. The result was a user-friendly spreadsheet with a concise summary of each resume and a binary qualification assessment for our HR team to use.
The automated system dramatically cut down the time spent on initial resume evaluations, enabling HR staff to focus on promising candidates.
The large language models provided a reliable 'pass/fail' assessment that was in line with the evaluations of the HR team, maintaining the firm's standards in candidate selection.
The new system can comfortably handle an increasing volume of applications, ensuring the firm is well-equipped to manage future recruitment surges.
The data preprocessing step ensured the privacy of applicants, removing personally identifiable information from resumes before assessment.
We demonstrated an innovative approach to HR challenges, setting a precedent for further AI applications in our recruitment process.