Machine Learning Engineer Offline Exercise

SalesLoft is looking to deploy a model to production that determines the seniority of a person based on their job title. This offline exercise will demonstrate some of your abilities related to this project.

Requirements

• Python 3.5 or greater
• Installed dependencies (pip install -r requirements.txt)
• A SalesLoft API Key (the recruiter will provide this)
• Training data (data/title_data_for_model_fitting.csv)

Getting Started

Copy (do not fork) this repository and follow the tasks listed below. Upon completion, please commit the code to a new GitHub repository or zip up the files and share with the recruiter.

Your Task

In exercise/model.py you will find the SeniorityModel class. You can train this model using the data found in data/title_date_for_model_fitting.csv. Your job is to follow the below tasks which further enhance the capabilities of the model class.

1. Implement a predict(job_titles) class method that accepts an array of job title strings and returns the predicted seniorities.
2. Implement a predict_salesloft_team() class method that loads all people in a team via the SalesLoft API and returns an array of tuples: (id, seniority).
3. Implement a save(filename) class method which persists the information in the SeniorityModel to disk at the given location. The model should be saved in a language-agnostic format (i.e. don't use pickle) and contains all the information that is required to create a new instance of SeniorityModel from the saved data. Assume that someone could use this file to construct an equivalent model in another programming language, given documentation.
4. Implement a load(filename) class method that loads a saved model (see above) from the given location on disk and initializes the instance of SeniorityMethod using the information in the file so that it can be used for predictions without fitting.
5. Use a testing framework to assert that your save/load functions are working. Make sure that a set of predictions from the persisted model match that of the hydrated model.