Gojek Takehome Assignment

As a senior data scientist at Gojek, you will often find yourself reviewing the work of junior data scientists. You will be expected to provide mentorship and guidance on best practices and pitfalls. You should adopt a similar mindset when completing this assignment.

This assignment is designed to be completed in 2 hours.

Getting Started

You will need to set up your development environment using conda, which you can install directly or via pyenv. Then, ensure you have GNU Make installed before running the following command.

make setup_env

The code is designed around several scripts that simulate a typical machine learning workflow. You will need to review and fix various issues pertaining to data cleaning, feature engineering and model training. Upon successful installation of the required packages, please proceed to run the pipeline using the following commands.

make data
make features
make train
make predict

You will run into errors along the way. Fix them, and proceed to run the pipeline again. Remember, errors in machine learning are often silent and difficult to catch. Spotting them will be key to succeeding at this assignment.

Submission

After you have finished fixing the pipeline, ensure that your pipeline works from end-to-end by running the following command.

make run

You should generate the following submission files. First, metrics.json containing model evaluation metrics.

{
    "your_metric_here": 0.5,
    "your_other_metric": 0.8
}

And also, results.csv containing your predictions of the best drivers to allocate.

order_id	driver_id
100032007	987454392
100167816	790186080
:	:

Be sure to fix any tests that have failed before submitting your code.