This project aims to address the limitations of traditional ATS (Applicant Tracking System) scanners while providing a more user-friendly solution for job seekers and hiring teams. It compares the candidate's resume with the specific job requirements and background fit needs and provides a more accurate analysis of the candidate's strengths and relevancy with the job requirement.

How it Works:

CVCompass utilizes a pre-trained LLM fine-tuned on a dataset of resumes and job descriptions. When provided with a resume and job description, the model performs the following steps:

1. Text Processing: Cleans and processes the resume and job description text.
2. Keyword Extraction: Identifies key terms and phrases in both documents.
3. Matching and Scoring: Matches keywords, assesses experience and project relevance, and calculates the ATS score and related metrics.
4. Feedback Generation: Creates specific feedback points for resume improvement.
5. Store a copy of feedback for fine tuneing model purpose and send the information over to client.

Future Development:

1. Improved Contextual Understanding: Enhancements on model by fine tuneing with custom data to better capture the context of skills and experience to generate optimal results.
2. More Model Options: Expanding the available model options to include more open source LLMs( fine-tuning options).
3. Recruiter View: Implement multiple resume evaluation at a time.

The Diagram below shows the workflow of how CVCompass works ->

Have any idea or recommandation? You can reach out to me at: iam.paulsayantan06@gmail.com

Star this project on Github :

1. https://github.com/SayantanmPaul/cv-compass-client
2. https://github.com/SayantanmPaul/cv-compass-server