October 14, 2015
Google had it right from the beginning when the tech giant took advantage of its massive asset – data. Google is in the business of data monetization and it wasn’t long until everyone else caught on to the game changing strategy. The industry we’re in, banks, as well as other industries with legacy data and systems, are also realizing the true power of data. With this explosion of data seeping through every crevice of nearly every industry, there came a need for someone to digest, analyze and make sense of it all. Enter the data scientist.
“It’s not rocket science” – unless you’re hiring a data scientist. Hiring a data scientist today that’s a good fit for your company – the right background, experience, education, skills, and personality – is so complex like rocket science, it seems almost impossible. Any HR exec will tell you it’s difficult to place the perfect person for any job, but data scientists are a different breed. The dataverse is experiencing explosive growth, and with it comes the need for this new knowledge set, but because it’s a relatively new field, every organization is clamoring to fill these positions.
And that’s not all. Specifically, skilled data scientists need to be highly knowledgeable business workers and technologists who can mine big data for key insights or steer a machine learning project – and that exact skill set is incredibly hard to find. Machine learning, while it may seem like a familiar buzz word, is still in its infancy, and data scientists need to have the creative drive to push it forward, which adds another complex layer to the mix of qualifications. But once a jewel of a data scientist is found who holds these traits, the word gets out very quickly and no employer who hires one is safe. Demand is hot and you better believe if you have a data scientist on your payroll, they’re getting job offers on a daily basis.
So, the supply is low and the demand is high. Here’s how to make your search less painful than it has been. At Feedzai, we’ve been going through the same process – here’s what we’ve learned along the way.
Choose your battles. Hint: think long term.
Before you gear up for the hunt, realistically decide on what skills are critical and optional, and make sure everyone on the hiring team is on board. This will define your pool of candidates and keep everyone on the same page when a candidate is up for consideration.
For example, with us, a good understanding of machine learning models and communication skills are critical because these are much harder to teach and foster. You must have teamwork and strong communication skills.
If the candidate isn’t experienced or as well versed in specific programming languages like python coding, we are more likely to look beyond it because, with the right candidate, that skill is teachable and it’s a good use of our time to invest in developing this scientist if they have strong qualifications.
Don’t look on a general hiring site.
If you think you’re going to find the needle in the haystack, you need to dig pretty deep and strategically. I wouldn’t dismiss taking a quick gander on a site like Monster or CareerBuilder, but you want to be where the data scientists are, and they are at academic institutions and in expert communities. Go to hackathons and spot folks who made people’s heads turn with their implementation
ns, or work your network at trade shows and events. Sometimes, you don’t even have to look that far; some of the best people we’ve found were right under our nose. Ask your top-notch colleagues if they know anyone because smart people tend to keep the same company.
Cut your search and interview time in half.
The initial search can be done with simple things like weeding through your digital resumes by plugging in key terms that represent what skills you want. In our case, we look for expertise in Cassandra, machine learning algorithms, Hadoop, Spark, scripting languages such as Python Coding, testing and training models with R, Weka, Matlab, and more. It’s likely you’ve got a big stack of resumes so put the ones who don’t mention those specific technical skills to the side. It doesn’t mean you won’t need them later but you should speak to the folks who have those technical skills you’re looking for first.
When it’s time for a candidate to come in, do not, by any means, conduct this interview process like you would with a business executive. This is a very unique and critical role in shaping your business so tread very carefully. Have your team ask very specific questions that root in problem solving, communication skills, technical and analytical rigor.
Then trust your team’s instincts. Do a few people have lukewarm feelings about a candidate’s performance or qualifications? Not sure how they will fit into the company’s culture? Then your candidate is likely a no.
Create a challenge; if you don’t have time to create it on your own, there are many sites that offer this. We’ve done tests asking candidates to predict how many people can survive the Titanic. Candidates deliver a daily project plan in which you can analyze how their minds work and how they view data in new and abstract ways. Do not forget to make candidates present their findings. Data scientists need to have effective communication skills because their findings will affect various teams within your organization and they likely won’t be technical fields. Data scientists are only as good as the team who is able to take their findings and implement them.
It’s not them. It’s you.
I considered talking about this earlier because it’s so important. Because data scientists are expensive and always in-demand, assume that the candidate is speaking to a few other companies during your interview process. Keep in mind, according to a recent survey, the median salary for an entry-level data scientist on the West Coast is $110,000. Now put this into perspective of where you live in the country, the cost of living, and nearby business demand for this role. Come up with a salary that’s enticing enough that a data scientist will want to invest his or her time in your organization. Also create a career development program that entices candidates to want to join your company – and keep it alive and well. Often times, I’ve seen companies say they enrich and enhance careers for its employees. Avoid the smoke and mirrors and actually make it happen. They can sniff B.S. from a mile away. For example, you may want to offer the ability to enroll in classes or continuing education programs at a local university.
And, lastly, like data science, your hiring process needs to always be re-evaluated for improvement.
These tips are what worked for us and I’ve talked to quite a few others who agree, but it may not always be the case for every company. If your process to find the perfect data scientist isn’t working, do what data scientists do—iterate and try something else. Talk to your peers at other companies on what works for them. Even better, find out what didn’t work for them and learn from it.
Did you like this article?
Get more delivered to your inbox just like it!