
This project is an AI-driven video screening platform designed to help recruitment companies dramatically speed up their hiring process. Recruitment companies create screening campaigns by uploading a job description. The system uses AI to automatically generate relevant interview questions, which recruiters can review, edit, and customize. For each question, recruiters can define desired keywords and evaluation criteria. Once the screening is ready, the platform generates a unique interview link that can be shared with candidates. Candidates open the link, review the job details, and record timed video answers for each question directly in the browser. Each response is captured as a separate video and uploaded automatically. On the backend, uploaded videos are processed asynchronously using a queue-based system. Background workers convert videos to standardized formats, extract audio, and store all media files. AI services then transcribe the audio, evaluate the candidate’s answers against the interview questions and desired keywords, and analyze video signals such as confidence, engagement, and potential cheating indicators. All results are stored in a database and presented to recruiters through a dashboard that shows detailed per-question analysis, transcripts, scores, and an overall ranking of candidates. This allows recruiters to quickly identify top applicants without manually reviewing every interview. The entire system is fully containerized using Docker and Docker Compose, allowing it to be run locally with a single command for development and testing.
7 Feb 2026