Gemjob is a web application to support clients on the freelancer platform Upwork in writing and maintaining job postings for data science and analytics jobs.

Support for clients is provided by suggesting recommended values for the various fields contained in the job posting form. These suggestions are based on similar jobs identified through clustering and the k-nearest neighbors algorithm. Additionally, we trained regression and classification models for suitable fields. Those include the job type (fixed price vs. hourly), the budget value and the feedback value.

The server and machine learning code is written in Python and hosted using Docker containers.

The images below show the dashboard, real-time recommendations, detailed predictions based on our models and our modular architecture (from left to right).