Cloudera Applied Machine Learning Prototype Hackathon

Phase 1
During this phase, teams will have 4 weeks to develop and deliver a proposal for a new AMP.

Your proposal must include:

• Problem Statement: what is the specific problem that your AMP will be solving
• Approach: how will your AMP employ ML methods to solve this problem
• Datasets: what open source datasets will be leveraged (provide links)
• Differentiator: how will your AMP stand out from the other submissions

Phase 2
For those selected to move on to the project phase, you will have 2 weeks to work on and provide 2 main deliverables. (1) a GitHub repository structured as outlined below, and (2) a video, no longer than 5 minutes, that explains your AMP to the judges.

Your GitHub repository should adhere to the following file structure:

├── src/        
├── scripts/
│   ├── download_data.py
│   ├── install_dependencies.py
│   ├── train_model.py
│   └── launch_app.py
├── app/
├── static/
├── .project-metadata.yaml
├── README.md
└── requirements.txt

File structure explanations and examples:

• src/ - Core code/classes necessary for the project (this is code that generally defines processes)
• scripts - Utility scripts for project and application setup (this is code that generally executes processes, these scripts will be called by the .project-metadata.yaml file)
• download_data.py - script to download necessary dataset
• install_dependencies.py - script to install dependencies pinned in requirements
• train.py - script to prep data and train and/or validate model
• launch_app.py - script to launch the visualization/application. The linked example is for a Streamlit application, though several other options and examples exist, just take a look in the AMP catalog.
• app/ - Assets needed to support the front end application (this is code that purely supports the front-end). The linked example goes along with the launch_app.py example also linked, so it will be a good example for a Streamlit application.
• static/ - Any images referenced in project docs
• .project-metadata.yaml - Declarative configuration file that specifies necessary steps to recreate the project in CML. The linked example includes comments for specific lines that may need to be modified. If you follow the exact naming convention documented here, then you will not have very much that needs to be modified.
• README.md - Should fully describe your project including: overview of your AMP, a brief overview of any novel ML techniques used, an explanation of the project structure, and instructions to run your AMP.
• Requirements.txt - list of pinned dependencies