MyDish

Cross Team project

This was a team project that continued to evolve as new teams joined and left. In my time on the project, my team was able to:

• Train an LSTM prediction model using Tensorflow/Keras to predict recipe ingredients in efforts to ease recipe entry and increase user engagement and retention
• Deployed an EC2 instance for use in iOS & Android app
• Implemented PostgreSQL database to log accuracy in efforts to refine prediction model
• Invoked Google Vision API for processing printed or hand-written recipes from user’s phone camera
• Deployed AWS Lambda function to parse GV return for recipe recording

Style Transfer

Using public domain images to map specific artists onto photographs

A sample of a few of the generated images

Song Recommender

Exploring multiple recommendation algorithms

This was a cross functional team project that has since been taken offline. As of June I have started to rebuild this project expanding it for better data visualization and explanations of how the recommender works "under the hood".