CV Recommendation 👨🏼‍💻

The goal of this project is to recommend top 5 CVs based on the job descriptions of the companies.

The archieve folder contains all the pdfs of the candidates. The folder itself contins a folder named data which again contains another folder called data. The child folder contains different other folders each named according to certain domains e.g ENGINEERING, BUSINESS-DEVELOPMENT etc. They contains the pdfs of the respective candidates.

How to run the application:

Fork the repository:

git clone https://github.com/asifrahaman13/cv-recommendations.git

Run The Scripts

Create a virtual environment. It is not madnatory but highly recommended.✅

virtualenv venv

Activate the virtual environment.

source venv/bin/activate

Install necessary packages in you environment.

pip install -r requirements.txt

Use this activated kernel to run the index.ipynb file. You can run the index.py file alternatively.

There are two executable python scripts namely index.ipynb, and index.py files. The index.ipynb file was used for building the primary code structure while the index.py file consists of class based components. The codes and architecture of both are similar. The index.py file is more structural and consists of more reusable components.

To run the index.ipynb file just click on the run all button on the top of VS code or Jupyter Notebook.

You can also do it through command line.

cd codes/
jupyter nbconvert --to html index.ipynb

To run the index.py file for the unix based system:

cd codes/
python3 index.py

For others:

cd codes/
python index.py

Run Docker Container

If you want to run the docker image then use the following steps:

docker pull asifrdocker/cvrecommendation:tag

Next run the docker container.

docker run asifrdocker/cvrecommendation:tag

Additionally to find whether the container is running you can run the following commands to check the logs:

docker ps

Copy the container id

docker logs -f <container id>

⚠️Note:: The code is written utilizing tensors. If your system has GPU Support (at least 4 GB of GPU Memory) then you will get good acceraleration in your computations. However if your device do not have GPU access the program will still run but at much slower rate. This is due to the fact that there is a lot of data and Python will reqauire some time for computations. If you find any error related to GUP out of memory please close your vs code and then re run the codes. 🐍 However still it may run out of the GPU memory. In that case please run the code in google colab or comment out the following line:

device = torch.device("cuda" if torch.cuda.is_available() else "cpu")

Next use the following instead:

device = torch.device("cpu")

✅For reference: In my Hexacore Rynzen 5 5600H processor with 4GB RTX 3050 Graphics card and 16 GB RAM, it took approx ~5 min to execute the entire code. It took ~3 min to extract all pdf data. ~2 min to perform the similarity and other preprocessing operations.

Approach:

First I extracted all the required information fro the job description as well as the CV data.
Next I tried to find the domain for which the companies are looking for using simialrity score. Similarity score is calculated between the job title of the company job description and the available catefories of the CVs. Only the CVs from the relevant categories are selected.
Next I used cosine similarity score between the job description and the CV data. The CVs are ranked accordingly.

Challenges face:

The most difficult challenge faced was based on the time of execution. I used the following techique to solve the problem:

Removed the CV directories to the archieve folder removing the intermediate folders.
Used parallel processing utilizing all the CPU units available. This helped to complete the execution time to nearly ~ 5 minute from ~ 26 minutes. About ~ 5 times more efficient. 🔥🔥

MIT License

Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the "Software"), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions:

The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software.

THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.

asifrahaman13 / cv-recommendations Goto Github PK