- AppliCruiter is an application powered by machine learning that streamlines the resume screening process by efficiently categorizing and assessing resumes submitted for job openings.
- Refactor UI to be more user friendly and intuitive
- Add authentication guard to routes to prevent unauthorized access
- Create new models for individual organizations and refactor existing models accordingly
- Refactor routes to accomodate for data changes
- Auth0 token persistence and expiry verification
- Incorporate new changes into the client (In Progress)
- Provision users (other than owner) to access different organizations
- Modularize and encapsulate components that can be re-used
- Implement proper data store using NGRX instead of relying on simple Subscriptions
- Develop e2e tests using Cypress in order to verify proper functionality of the application (In Progress)
- Written in Python, making use of the sentence_transformers library integrated with PyTorch. To enhance performance, we fine-tuned the all-mpnet-base-v2 model using an extensive dataset of resumes. Furthermore, we incorporated the pdfplumber library to effortlessly extract text from PDF resumes.
- A MySQL database was leveraged for storing persistent data related to jobs and resumes. This relational database ensures efficient and organized management of information crucial for seamless operations.
- Amazon AWS S3 was leveraged to store resumes in the ubiquitous PDF format. Leveraging S3 buckets provides a scalable and secure solution for efficiently managing and accessing resume data, offering a reliable infrastructure for storage needs.
- Developed REST server using Flask to serve as the bridge that seamlessly connected our model and persistent data to the client. This permitted the creation of a responsive and efficient web application that can handle requests, process data, and deliver results.
- Built With
- Python
- PyTorch (sentence_transformers)
- Flask
- AWS S3
- MYSQL
- all-mpnet-base-v2 (fine-tuned with a comprehensive dataset of resumes)
- pdfplumbler
- JWT
- Written in TypeScript, making use of Angular to create a user-friendly interface for the application.
- The client seamlessly interacts with the server, providing real-time updates for each job posting created.The use of RX.js allowed observability, which made the application dynamically responsive to state changes. Whenever a new job posting, a new resume is created, the interface automatically updates in real-time, providing an efficient and intuitive user experience.
- Cypress was utilized for end-to-end testing to ensure the seamless functionality of the application from the user's perspective.
- Built With
- TypeScript
- Angular
- Material UI
- Tailwind CSS
- Cypress
- Auth0
- Client
- Node - 18.17.0
- Server
- Python - 3.8.18 (MiniConda - Python Environment Manager)
- MYSQL - 8.1.0
- S3 Bucket
- Navigate to client directory within the project
- Verify that you have the correct version of Node installed
- Run the following:
npm i - Manage the environment variables to properly utilize your hosted server
- Navigate to the server directory within the project
- Make sure your python virtual environment is currently activated
- Install required packages and dependencies with
pip install -r requirements.txt - Externally make sure you have setup an S3 bucket to store data and that MYSQL is currently running
- Populate a new .env file by referencing the .env.example file
- Download the model from the dropbox in the resource section and unzip directly in server directory
- Run truncate_all.py to configure the MYSQL database and truncate the data within the given S3 bucket
python truncate_all.py - Start the server with
python main.py - Start the client with
npm start
Feel free to shoot me a message or connect with me on LinkedIn!