The idea of a job description is simple: give information about the job itself, describe the responsibilities, and state what kind of person you are searching for including the skills. Over the years when I was surfing LinkedIn for job posts I came across quite a lot of job descriptions that made me turn right away. I can see the same behavior in the community of data scientists and machine learning engineers. In this post, I’ve listed five examples that made job descriptions a repellent, instead of an attractor.
- Don’t Invent Your Own Job Titles
It might sound cool or interesting to post a job title that is the product of your own imagination. Maybe you think that it describes what you expect perfectly, or will make it look interesting for candidates. It is quite hard to describe a profession with a couple of words. There is still a lot of discussion around the term “data scientist”. If you don’t really know what you are doing, take a safe bet. Don’t invent your own job titles. Invented job titles don’t only make it less appealing, it makes it a fun stock. Our data science slack channel was full of these kinds of job titles where we made fun of them.
2. Don’t Ask the Impossible
Usually, this is an issue about the expectations, which projects to job descriptions. If you are willing to hire a technical person, you should have a sense of what is possible in that area. This doesn’t mean digging into the technical details. Nevertheless, you should have a feeling of what is possible with current technology and what is not. It is true that AI can unlock new possibilities that were not possible a short time ago, however it cannot unlock every door.
3. Don’t Focus On Specific Models/Algorithms
There are a lot of methods in machine learning. You probably heard many: random forests, neural networks, linear models. The list just goes on and on. Surely there are more and less popular ones. The last time I checked neural networks (although a quite broad term) and random forests were one of the most popular methods listed in job descriptions.
Machine learning practitioners usually don’t focus on one particular method. Instead, they discover and learn new methods for different problems continuously. Listing specific methods in the job description will not help you to find a better candidate. Note that an experienced machine learning practitioner can learn a new method easily by simply reading its paper. Check out this post of mine for a more detailed discussion.
usually don’t focus on one particular method. Instead, they discover and learn new methods for different problems continuously. Listing specific methods in the job description will not help you to find a better candidate. Note that an experienced machine learning practitioner can learn a new method easily by simply reading its paper. Check out this post of mine for a more detailed discussion.
4. Use Correct Terminology
It is understandable to become confused with terminology if you don’t have an education or experience in this field. However, this is not an excuse to do a better job. It is quite disappointing to see the wrong terminology being used in the wrong places in technical job descriptions. It gives a wrong signal to candidates.
Deep learning and artificial neural networks are not two different things. SQL is not a programming language. Moreover, it is quite not clever to ask your candidates “how many libraries” they know in Python. Do your own research or ask for help. Else, soon you might be asked to distinguish between pokemon and libraries:
5. Search for Skills, not Number of Years
Asking for “experience” for a number of years is a bad shortcut. Doing the same repetitive task for a number of years doesn’t mean experience. Moreover, working for a company is not the only (and probably not the most effective) way to acquire knowledge and experience in a field. This is especially true in computer science related fields like data science and machine learning. Instead of focusing on experience, try to understand the skills of the candidates.
Understanding the technical skills of a candidate is hard. Especially for non-technical people. Elify.io is a skill assessment platform for data and machine intelligence job suites. It is designed for non-technical people, by experts in the field; makes it easy to understand the technical skill levels of candidates.
Ben je op zoek naar advies? Klik hier om in gesprek te gaan.
