AI Recruiter

An AI based intelligent way to find out best among the crowd

This is the source code of our BTech final year Project

Team members -

• Tarun Agarwal
• Sandeep Kumar Shukla
• Sahyog Saini

Steps to get backend running

1. git clone <repo>
2. pip install virtualenv
   python3 -m venv aienv (for linux)
   virtualenv aienv (for windows)
3. pip install -r requirements.txt
4. python manage.py migrate
5. python manage.py createsuper
6. python manage.py runserver

Resume similarity score calculation -

Metrics considered -

• Jaccard Index
• Sorsen's Index
• Cosine Similarity
• Tversky Index

Given the parsed resumes, we are selecting a subset of fields to calculate suitability score.

Resume fields we parse are -

• name
• email
• education (do not consider because it just increase number of tokens, thus decresing score)
• technology (do not consider since not present in every parsed resume)
• skills (consider because of obvious reason)
• experience (consider because of obvious reason)
• projects (consider because of obvious reason)
• profiles (not useful at all)
• positions (not present in all parsed resumes)
• interests (are primarily non-tech, hence not useful)
• certification (can be achieved easily, hence is not important)
• awards (irrelevant)