Vizy

Visa Process Made Easy

Try Now

The application is a mock visa interview platform where users can give interviews. The system utilizes machine learning to analyse the interviews and responds with the probability of success.

Vizy Backend

Node js application frontend to integrate UI with backend.

a) Integrates with chatfuel api's to generate interview questions for the user from a large of questions scrapped online.

b) Gold standard training data is taken from good websites where users have submitted their interview experiences.

c) Trained and Tuned RandomForest model finds the probability of success of the user using extracted features - Glove 300 dimensional glove embeddings - GRE/TOEFL scores - Length of answers - Sentiment of answers - Consulate

d) User is notified using Twilio api's when interview feedback is ready.

e) User interviews are then uploaded to Dropbox for processing. Model picks up the interview files and responds with detailed response.

Getting Started

Deploying nodejs application is done using

npm start

Deployment

The application is deployed on heroku.com

Built With

• Nodejs - The web framework used
• Chatfuel - Chatbot workflow
• Google Analytics - User Analysis
• Facebook Customer Chat - Social Login

Built With

• css
• dropbox
• html
• javascript
• machine-learning
• natural-language-processing
• node.js
• python
• scraping