Consider Unique Business Needs When Hiring a Data Scientist
Many companies hire data scientists in the information economy, but the profile of these candidates is changing rapidly. It is not that there is no place for the doctorate. statisticians or mathematicians who are also R or Python coders, there is not enough for everyone. Another problem is that not all of these brains have great people skills or the ability to solve business problems.
Does that mean companies should just hire someone who calls themselves a data scientist? Certainly not, but companies are wise to think about their actual talent needs, as these can differ depending on whether a company is hiring its first data scientist or joining an existing team.
Qualification requirements have changed
When the Executive Director of the Deloitte AI Institute, Beena Ammanath, worked for E-Trade, there were data scientists (called analysts or statisticians at the time) doing all the data predictions and simulations; with a BI team that produced historical reports and an ETL team. When Hadoop and machine learning started to become popular, organizations sought out data scientists with PhDs who were responsible for collecting, cleaning, and preparing data, as well as setting up tables.
“They were horrible doing it, so you could never expand it beyond the lab environment,” Ammanath said. “We always see companies with data scientists doing 12 to 15 different jobs, but the more mature companies separate them into data engineers, data visualization, QA and test engineers, machine learning engineers. [and] MLOps engineers. “
This fragmentation of the role of data science is actually a good thing, given the shortage of high-level data scientists. Also, Ph.D. Data scientists are expensive, so it is more economical to employ a range of positions.
Meanwhile, data science tools have become easier to use, allowing more people to connect to data sources, prepare data, build models, and analyze data.
“If you look at the existing programs offering data science degrees, they don’t necessarily agree on the qualifications,” said Alin Deutsch, chief scientist at TigerGraph, a database and platform provider for data science. graphical analysis and director of the data science program at the University of California, San Diego (UCSD). “TO [TigerGraph and UCSD] we want data scientists to have a foundation in several different fields that involve understanding the principles of those bases. “
Understanding the principles of data science is important because languages and tools evolve so rapidly that everything data scientists use today might not be what they will use tomorrow. If a data scientist has a solid background in statistics, even an undergraduate degree may be sufficient.
However, if they are hiring a data manager or an AI manager, organizations should consider the differences in roles.
“A Chief Data Scientist is someone who has a PhD, has deep knowledge of technology, does hands-on coding, and knows the latest and greatest in AI research,” said Ammanath. “An AI manager is someone who understands IT and has an MBA. “
When hiring a data scientist, Deutsch seeks a mastery of object-oriented language, experience in statistical and probabilistic applications, knowledge of certain machine learning techniques and commonly used algorithms such as clustering and linear regression. . Ideally, candidates should also have experience visualizing large data sets and presenting them to non-expert consumers, he added.
“They shouldn’t be learning these on the job,” Deutsch said. “As to how to apply them, ideally it’s in an end-to-end pipeline where they had to start with data that wasn’t clean, clean it up, combine it, run machine learning algorithms on it and pull it off. conclusions based on their statistical knowledge. . “
Applicants earn bonus points for being exposed to relational databases and NoSQL. Specifically, the experience or the ability to speed up a query on large data sets by instituting clues and pre-materializing the computation.
Alternatively, if you are hiring a recently graduated data scientist, look for relevant school or extracurricular projects.
“If I had to give someone in school a piece of advice today, I would say, ‘If you’re doing a project at school and you want to be a data scientist, pick a project where the data is. the heart of the project, like building a recommendation system where you really have to deal with the complexity of the data, ”said Ira Cohen, chief data scientist at Anodot, an AI platform provider that focuses on anomaly detection for business monitoring.
Checking the fit
Organizations have different ways of testing data scientists to determine if they are qualified for the job. For example, Cohen provides a candidate with a business problem, data extracts, and the desired outcome.
“I want to hear first that I’ll do A, then I’ll do B,” Cohen said. “Then I say, ‘OK, you said A, describe it.’ If you know what data science is and you know how to do it, that’s natural. “
Deutsch proposes a problem to be solved asynchronously: the candidate brings the problem home and then comes back with a solution.
“You could take a small project in that direction – clean it up, learn from it, give me a conclusion,” Deutsch said.
Depending on what they use, some organizations test knowledge of specific languages such as Python or R, or the candidate’s experience with using technologies such as Snowflake or Spark. However, it’s important to remember that a tool-only test is probably short-sighted, given the speed at which languages, technologies, and tools change.
Fundamentally, problem-solving skills are essential regardless of a candidate’s training and experience, as some of the brightest minds can work magic in a lab that doesn’t translate well into production. What businesses really need are data scientists who can help them achieve their business goals.