Virtual-Mock-Interview

"Virtual Mock Interview" or VMI is a platform that helps job seekers enhance their interviewing skills and helps companies filter potential employees using multiple ML models to analyze the interviewer behavior and their ability to answer questions.

Previews

Models and their usage:

Facial Models:

• Mediapipe for iris tracking --> Indicates focus level
• Residual Masking Network for facial emotion detection --> indicates energy level

Audio Models:

• Librosa a deep learning audio analysis library
• whisper for the speech to text task

Large Language Model:

• GPT API for feedback in a natural language form.

Configuration:

Backend

1. Clone the project
2. Navigate to the backend directory cd virtual_mock_interview/vmi-api
3. Setting up conda environment:
   • Download Anaconda if not installed
   • Create conda environment conda create -n myenv python
   • Activate it conda activate myenv
4. Download required dependencies pip install -r requirments.txt
5. Setting up mysql database for questions table:
   • Download mysql workbench` and mysql server if not installed through download link (Note: check 'start mysql server at system startup' and 'grant mysql server file permission' to avoid tons of hustle later. Watch a tutorial if confused )
   • Start a new mysql connection and fill the information (Note: You can fill it anything but make sure you remember them as you will enter them in .flaskenv later on)

      connection name: vmi-schema
      host name: 127.0.0.1
      port: 3306
      username: admin
      password: #set your own password
     

     then test the connection, if successful press ok and click on vmi_schema
   • Right click on tables, table data import wizard and add the questions.csv file which could be found here questions.csv
6. locate .flaskenv_example in the backend directory, rename it to .flaskenv, and add your database credentials. You will also need an OpenAI API key which could be optained from API Keys
7. Initing the flask server and the local file server:
   • Open a terminal and navigate to the backend directory cd virtual_mock_interview/vmi-api
   • Activate conda environment conda activate myenv
   • Run the flask server flask run (Note: the first run will take a while as both models whisper and Residual Masking Network download and cache their model weights)
   • Once the flask server is up and running open another terminal, navigate to the backend directory and run the local file server python local_file_server.py

Frontend

8. • Open a third terminal, navigate to the frontend directory cd virtual_mock_interview and run npm install to install the frontend dependencies
   • Start the npm server via npm start

Documentation

Detailed documentation could be found in docs

Issues or inquiries

Please reach out if you have any question regarding this project by openning an issue, DMing one of the three contributers or email me at [email protected]

Ahmed Aly.