saish-vc/resume-screener-app

This is a local web app that helps recruiters and hiring managers sort through resumes. You upload resumes in PDF, Word, or plain text format, and the app pulls out contact details like email and phone number, compares each resume against a job description you provide, and scores how well each candidate matches. The matching score combines two things. Seventy percent of the score comes from comparing the overall wording and themes of the resume against the job description, using a common text similarity technique. The other thirty percent comes from checking a resume against a built in list of over sixty programming languages, tools, and skills, and measuring what fraction of the skills mentioned in the job description are also found in the resume. Based on the combined score, each candidate is labeled a high, medium, or low match, using preset percentage cutoffs. The app is built with Streamlit, a tool for quickly building web interfaces in Python, and shows a sidebar where you can upload multiple resumes at once, or try a demo mode that loads sample resumes automatically instead. It highlights which required skills a candidate has and which ones are missing, lets you inspect the raw text extracted from any resume, and gives an overview of the whole applicant pool, including the average match score and which skills are most commonly missing across all candidates. Results can be exported as a CSV file with each candidate's score, contact details, and match verdict. A companion script can generate a handful of realistic sample resumes for testing, so you can try the tool immediately without uploading real resumes of your own. Getting started only requires Python 3.8 or newer, installing a short list of dependencies, and running one command to start the local server, which opens the app in your browser. The README describes the project as an early, minimum viable version, focused on giving a working demonstration of automated resume screening rather than a finished product.