Data science hiring is a reasonably tricky task as hiring for it without understanding the skills, tools, and technical expertise they possess will lengthen the process. Not just theoretical but practical experience of the tools, but the ability to build solutions and real-world use cases matter the most while hiring data scientists. Additionally, with not much formal education available, some professionals might call themselves data scientists without having the right credentials, quickly becoming a grave challenge recruiters face these days.
Seth Dobrin, who heads IBM's Data Science Elite Team, has an excellent suggestion for recruiters. He suggests that if a company is building a data science team, the first step is to hire a Senior Data Scientist who can further lead to the team's development.
As the industry is still quite a niche, until senior professionals are on-board, it isn't easy to get others to come on board. Two years ago, Dobrin was hired to build out the Data Science Elite Team. In this new endeavor, IBM data scientists engage with organizations in six to 12-week engagements to collaborate on data science and AI projects. After spending a year traveling while meeting IBM clients, he successfully built a team of 60 data scientists, machine learning experts, and others with related expertise. Not just that, in 2019, he added 30 more data scientists to his team.
When hiring data scientists, large job directories, such as Glassdoor, Indeed, and LinkedIn, are very popular and are often the first choice for companies. Hiring data scientists typically includes getting applications, pre-screening, technical tests, in-person or virtual interviews, and selection. This can be a successful method; however, large tech companies avoid listing their job offers on these websites with a fear of getting too many applications. It is often difficult to find the right fit from a haystack.
Besides these, hiring data scientists through peer networks and external consultants is a good source. Given the talent pool is a niche, the employees might refer to friends, professional contacts, and acquaintances they know would fit a particular role. The field's nature is more research-oriented and unsaturated, so there is a high chance that professionals from this field are well connected.
Some smart ways of recruiting data scientists are also through non-traditional methods such as Hackathons, GitHub, Conferences, WhatsApp and Telegram Communities, and Local Meetups. You’ll find data science interview questions in the fifth section of the paper.
"In a competitive field like Data Science, strong candidates often receive three or more offers, so the success rates of hiring are typically below 50%. There is more than one way to source data science professionals; however, below are the three communities that stand out in efforts and outcomes."
Hackathons have become one of the popular methods in the analytics community to hire the right fit. Big and small, many companies are partnering with hackathon platforms to spot data science candidates. It is one of the top-rated platforms to demonstrate skills while competing with the best programmers in the domain.
They are a 24-48 hours event that provides an innovative and energetic environment where participants use different tools to analyze, visualize the outcome, and win the code race. Recently, many organizations have started collaborating and organizing hackathons to identify and gain new talent. Some also offer practice sessions where data science enthusiasts usually practice Machine Learning algorithms like Support Vector Machine (SVM), Linear Regression, Naive Bayes, Extreme Gradient Boosting Classification, and more.
Hackathons are one of the best mediums for sourcing data science candidates because they:
GitHub is one of the world's largest code hosts, with close to 50 million developers.
It is a perfect platform to showcase work by machine learning and data science enthusiasts. The platform allows collaboration with the team members to showcase coding skills while acting as an online resume. It is becoming a revolutionary platform to identify data scientists and their skills. Data science professionals use GitHub to host code repositories, data, and interactive explorations, present their work, and impress hiring managers. The job aspirants usually set up an account on GitHub to create a repository of their work.
The platform allows collaboration with the team members to showcase coding skills while acting as an online resume. It is becoming a revolutionary platform to identify data scientists and their skills.
StackOverflow is a Q&A site for professional and enthusiast programmers. Just like GitHub, StackOverflow is also an excellent platform to hire exceptional Data Science talent. It is a Q&A site where developers post and answer technical questions. Tech recruiters would need to carefully read the candidates' answers addressing specific questions to see if they are the right fit.
On StackOverflow platform, the developers are segregated based on their user badges and reputation scores. An ideal candidate ranks high for both, and that should be easier for recruiters to gauge. Every question posted has tags associated with it; they can be used to find users who fit the company's data science requirements. However, after connecting with a candidate, it is essential to validate the resume and conduct a tech skills assessment to shortlist him/her for the next round of interviews.
Some other places to find great data science talent is through machine learning challenges, similar to the hackathons we mentioned above. The coding challenges are great platforms for candidates to showcase their skills in action. While hiring top Data Science talent, testing candidates on real-time problem-solving skills can move up the recruitment efforts by days or weeks.
Are you hiring an entry-level data science professional or an experienced one?
Companies that do not have a reliable data infrastructure and internal BI practice need a data engineer first. S/he will build pipelines and prepare data for the data scientist to use. Many companies tend to skip this step because it's not the mainstream data science profile, but that is a mistake. If a data scientist is hired first, they won't have any data to work on, so they will either leave or deny working as a data engineer, as crunching data from scratch isn't something a data scientist does in his profile. Hence, the companies trying to establish a new data science team must hire a data engineer prior to a data scientist.
As mentioned in the section on "Building a hiring pipeline for Data Scientists," the companies need to hire a senior data scientist. Cutting costs or settling for lesser experienced data scientists won't help the company with problem-solving skills. They will move quickly with minimal assistance, giving the company a faster return on data science investment. The senior candidates usually command a higher salary than an entry-level candidate, but they typically are a revenue addition to the company. Hence, it's rather imperative that someone experienced steers the ship to the shore.