Back to article overview

Hiring Data Scientists - Best Practices & Job Description Template

by Juan Manuel Ortiz de Zarate in #allinhrresources #datascience

Hiring an IT candidate is one of the hardest tasks that human resource professionals have to accomplish. Demand for IT professionals is greater than available individuals on the market, which produces competition between companies for the scarcely qualified developers. This can result in a lot of poaching of talent from rival employers. Take, for example, the career paths of some of the most famous IT professionals Ian Goodfellow and Guido van Rossum.

Goodfellow is one of the world’s foremost experts in deep learning and one of the creators of Tensorflow, perhaps the most powerful neural network libraries. After getting his Ph.D. he started his career with the Google Brain research team, later leaving this position to join OpenAI, only to be hired again by Google a year later. Two years afterward he was employed by Apple to his current position, Director of Machine Learning.

The case of van Rossum is somewhat similar. As the creator of the Python programming language, he was lured into work by the biggest IT companies in the world. He first joined Google for seven years, then was convinced to work for Dropbox for six more years until retirement. Even after retiring, he continued receiving juicy job offers. Eventually, he accepted returning to work for Microsoft.

What is exemplified in these stories is that even IT giants, such as Google or Microsoft, have difficulties and must navigate a highly competitive environment to hire the needed professionals. Moreover, this challenge is even greater for finding experts in Data Science, because this field demands a stronger background in academic and professional experience. Therefore, it is paramount to be very convincing when attracting data scientists to a new job. Let's start a brief introduction to the profiles that a data scientist could have.

Need help attracting the correct talent? Tell us what you need! We can connect you with up to 5 companies to match your needs within 72 hours—for free!

Profiles

Although data scientists may have different academic backgrounds, such as physicists, economists, mathematicians, software engineers, or computer scientists, in practice they may have the same approach regardless of their study focuses. These approaches depend on their experience in the data science field.

Another aspect about their profile to consider is whether you should hire a data science freelancer or do you need to outsource a whole team.

Generalist

Data science is an extremely broad field, applicable to hundreds of different contexts. Each one implies specific knowledge, both of the problem in question and of the mathematical and computational theory to answer it.

A general data scientist is someone who knows a considerable amount about all theories, the classical tools to apply in each context, and how a good solution could be faced. This person has had many works about different topics and areas over his career. A representation of this expertise is Andrew Ng, who has completed works about: Natural Language Processing, Clustering, Deep learning, Images, Robots, Time series, and more.

This profile type is good for a person that leads a data science team, works on multiple projects (at the same time or sequentially), or must create completely new solutions combining techniques from different areas.

Specialist

On the other hand, there is a specialist, who knows everything about a specific area of Data Science. This type of candidate knows the latest and best tools and how to use them in this field. They are able to implement the best solution known and potentially further improve them.

For example, if you want to create a solution about Natural Language Processing, like a very sophisticated chatbot, a perfect candidate would be someone like Tomáš Mikolov. He is “famous” for working on NLP research for the last ten years and creating and implementing program frameworks able to understand free text and generate rational responses. However, for another kind of solution, like detecting objects with images, his implementation time will likely be too long and the solution won’t be very good. Most likely, he would reject such orders and search for other jobs more fitting to his skills.

The first thing you need to know is what kind of profile fits with client goals. Then, you can start your search and selection process without losing time in failed matching.

Hiring a Data Scientist

The first place many look to find Data Scientists is a classic job offer site, like Linkedin. These are not bad options, but there are more specific websites to gain better information about an applicant’s skills.

If you’re looking to recruit viable candidates, you may also be interested in our article about job description for data scientists.

Alternative sites to look for

Kaggle is one of the most popular sites for Data Scientists to show their skills. There are many different competitions for the creation of the best prediction machine learning models. Some competitions include high monetary rewards while others are motivated by the love of science. The site has a ranking section with all participating users listed, necessary contact information available in user profiles.
is one of the most popular sites to share findings, knowledge, and experiences in Data Science. It involves a highly active community and popular articles with contact information of authors available.
Medium is another very popular blog site, maybe more so than TowardsDataScience. It not only focuses on Data Science but has a section dedicated exclusively to it. Many articles and author contact information can be found there.
Last but not least, Pangea.ai of course! We are here to help you choose the right data science professionals, just tell us about your needs! We can connect you with up to 5 companies to match your needs within 72h completely for free!

Job Description Template

When you contact a candidate or post the job description on any job offer site, you should explain elements of the job such as commitment, job type (remote or on-site), company goals, task descriptions, technical skills or knowledge as well as technologies desired. If no technologies, knowledge, or skills are specifically desired, this must be stated additionally.

We provide you with a job description example, based on a job offer post of a popular international company, aiming to attract talent:

Job Description

We are seeking an experienced data scientist to assist us in discovering information hidden in vast amounts of data, make smarter decisions, and subsequently deliver stronger company products. Your primary focus will be in applying data mining techniques, performing statistical analysis, and building high-quality prediction systems to be integrated within company products.

Responsibilities

Select features while building and optimizing classifiers using machine learning techniques
Extend company data with third-party information sources, as needed
Processing, cleaning, and verifying integrity of data used for general analysis
Performing ad-hoc analysis
Presenting results in a clear manner

Skills and Qualifications

Strong understanding of machine learning techniques and algorithms, such as k-NN, Naive Bayes, SVM, and Random Forests.
Experience with common data science toolkits, such as R, Python, NumPy, and MatLab.
Great communication skills
Proficiency using query languages, such as SQL
Fair mastery over applied statistics skills, including distributions, statistical testing, regression, etc.
Good scripting and programming skills
Data-oriented personality

Additional information (optional):