AI Job Application Screener — Screen 50 CVs in 5 Minutes and Never Miss a Good Candidate
Built by EchoCipher ·
About this build
A Claude powered workflow that takes a job description and a batch of CV text and outputs a ranked shortlist with a one paragraph summary per candidate, a fit score out of 10, and the specific reasons why each person does or doesn't match the role.
What was built
Built this after helping a friend hire for a small team. She had 60 applications, a full time job to run, and genuinely no time to read them all carefully. She was about to just go through the first 20 and stop. The workflow takes two inputs. First, the job description pasted in full including any must have requirements. Second, each CV pasted one at a time or in a batch depending on how you structure the prompt. For each CV the output gives you five things. A fit score out of 10 based purely on how well the experience matches the role requirements. A one paragraph summary of the candidate written as if you were briefing someone before a phone call. A bullet list of green flags, things that are genuinely strong matches. A bullet list of yellow flags, things that need clarifying in an interview. A single line verdict — interview, maybe, or pass — with the main reason why. The scoring is not magic. It is consistent. And consistent is what you need when you are reading 50 CVs because your brain stops being consistent around number 20. What surprised me: the yellow flags section became the most useful part. It flagged things like employment gaps or vague job titles that I would have either missed or unconsciously penalised without realising why. Tested on 3 real hiring rounds across different roles — a marketing coordinator, a junior developer and a customer support lead. Shortlisting time went from roughly 3 to 4 hours down to under 45 minutes including reading the outputs. Happy to share the full prompt structure in comments if anyone wants to adapt it for their own hiring process.