What you build
A job portal where employers post jobs and candidates upload resumes. For each job, the system ranks applicants by how closely their resume matches the job description, and explains the match.
Modules
- Employer and candidate accounts
- Job posting and applications
- Resume text extraction
- Matching score using TF-IDF and cosine similarity
- Ranked applicant list with matched and missing skills
What you will understand
- Turning text into numbers with TF-IDF
- Measuring similarity between documents
- Why a matching score needs an explanation, not just a number
How it is tested
- A small labelled set of resumes to check whether the ranking makes sense
- Precision of the top-5 matches
- Test cases for empty or unreadable resumes
What goes in the report
- Literature on resume screening
- Algorithm explanation with a worked example
- Evaluation results and limitations
Viva questions to prepare
- Explain TF-IDF with an example.
- Why cosine similarity?
- How could this system be unfair to some candidates?