This project is to write out a proposal and document how you would build a system to continuously learn and enrich data related to job listings. To be clear, we are looking for a detailed proposal and plan along with an estimated cost or proposed project cost. We will review the proposal of several people that we invite to this posting and then based on the proposal build out the actual project to execute on their plan.
As you read through the rest of this, please feel free to message me on Upwork or on Skype. My skype address is amurphy59.
Specifically, we want to understand the recommended infrastructure and software (i.e. AWS EMR with Hadoop & R) necessary to execute a project that will deliver on the following project concept. We also would like to know what the development process is that you will follow
Our business is an online job search and recruiting service existing in many countries around the world.
Today, we have 110,000 active jobs, and over 500,000 total jobs (including those that have expired) to analyze from the past 90 days. On most days, there are about 10,000 to 20,000 new jobs added into our system. We are in the process of adding more and more jobs everyday. We expect the system when fully built out to have about 10,000,000 active jobs.
These jobs are often listed with a combination of attributes about the job in the advertised title. Attributes include the skills, compensation, the name of the company, and then the actual profession title itself. Each company that lists a job opening will often use their own category taxonomy.
In an ideal world, companies would adhere to a specific taxonomy, use normalized job titles, and list out skills from a specific skills catalog, but that doesn't happen.
The desired outcome of this system is to be able to create structure from these jobs listings, to enrich the data and identify the attributes about each listing at the point in time when the job comes into our system. To be able to take the learnings about past job listings and apply them to new ones.
Here are some examples of the desired output for each job record:
- A common profession title
- Information about the skills necessary for the job
- Information about the company
- Information about the benefits of the company
- Information about the requirements of the job itself
Each of the attributes above will be used to create a better description about the job listings when presented to job seekers.
Follow on projects will be set up to determine a similar set of attributes about the worker
execute similar logic with respect to resumes, and ultimately to make recommendations for both companies posting jobs and users posting resumes of where there is a great match.
In your response to this posting, please include the following:
1. Your proposed price. (Note, I have put a budget of $100 for this, but if you require more time and or a higher price, please include why in your response)
2. A min of two example projects using NLP / Machine Learning that you have participated in.
3. Your preferred software languages.
4. What you expect to produce in this write up in terms of work product.
Again, please let me know if you have any questions about this project either through Upwork messages or on Skype: amurphy59