pentakill-cv-analysis

This project was developed for the Experaince GenAI 2.0 Hackathon and focuses on CV analysis using Azure OpenAI. The goal is to create an automated system to rank and analyze CVs based on a job description by leveraging OpenAI, text embedding, and cosine similarity algorithms. The app folder contains the core components of the project.

Project Organization

├── data
│   ├── cv_pdf       <- input cv files
│   ├── jd_pdf        <- input job description files
│   └── ocr_text      <- extracted cv and jd .txt
│
├── app          <- Configuration file for flake8
│   ├── upload_cv        <- saved cv pdf files from file uploaded from html:js
│   ├── upload_jd        <- saved job description pdf files from file uploaded from html:js
│   ├── output       <- saved processed cv and jd in csv {name:ocr} files
│   ├── static       <- css and frontend media (img,png,mp4) 
│   ├── template        <- *.html 
│   ├── utils       <- functions called in routes.py
│   │   ├── files2csv.py        <- data processing from pdf to csv
│   │   └── gpt_openai.py      <- functions for retrieving and calling Azure openai 
│   ├── app,py       <- initialize flask app in local deployment (port:5000)
│   └── routes.py       <- receive HTTP methods from html:js and ex1cecute functions
│
├── requirements.txt   <- Dependencies (env:name:"cv-analysis")
│                         generated with `pip freeze > requirements.txt`
│
└── README.md          <- The top-level README for developers using this project.

---

cv-analysis with openai workflow

1. Processing multiple cv pdf and one job description pdf into csv files.
2. Text embedding cv_ocr and jd_ocr.
3. Using cosine similarity to calculate similarity score for CVs and one jd.
4. Ranking and filter out top 3 candidates respected to the jd.
5. Retrieve the ocr of the 3 candidates and 1 jd.
6. Combine [5.] and user input in to request prompt
7. Calling openai with prompt and config, return the response
8. Can continue communicate with openai and getting desire solution.

Step by step guidance:

steps for running the app. Creating python environment

• cd pentakill-cv-analysis
• conda create --name cv-analysis python=3.10
• conda activate cv-analysis
• pip install requirements.txt

Set up env config

• cd app
• create .env for the following environment parameters { AZURE_OPENAI_GPT4_ENDPOINT="" AZURE_OPENAI_GPT4o_ENDPOINT="" AZURE_OPENAI_TEXT_EMBED_ENDPOINT="" AZURE_OPENAI_API_KEY="" }

Then run the following code:

• flask run

Interface operation

• attached multiple resume.pdf files
• attached one job description.pdf file
• then click upload file
• then input prompt into chatbox

• for example: " I will provide you with a list of candidates' information and a job description. Your goal is to match the candidates with the job description and rank the candidates accordingly, with the most suitable candidate ranked first."

• then click "send"

Collaborators: Tan Beng Seh, Sui Wei En, Ng Rou Yan, Pah Onn Qi, Tan Yong Seng