IBM has tripled the size of its Data Science Elite Team, launched in 2018, to nearly 100 data scientists that have been deployed on more than 130 projects.
According to Seth Dobrin, Ph.D., vice president of IBM's Data and AI unit and chief data officer of IBM Cloud and cognitive software, the Data Science Elite Team works with software customers and operates independently from the services group. Typically, the Data Science Elite Team is brought in during the evaluation of IBM software or post-sale.
IBM's Data Science Elite Team works with software customers to speed along machine learning and artificial intelligence implementations based on methodology honed through enterprise deployments. Design thinking, executing on sprints, and business knowhow leave the client with a blueprint to follow.
Here's what you need to know about how IBM structures its Data Science Elite Team and what it looks for in recruits.
Composition: Dobrin said that the team is comprised of 25% early professional, 25% highly experienced, and the remainder mid-career or later. Of IBM's Data Science Elite Team hires to date, 15% have been internal candidates with the remainder external.
Expertise needed: The one common thread of all the roles on the data science team is Python. "The most basic requirement is having deep expertise in Python. That covers all job segments," said Dobrin. Other languages that matter depending on the role include Scala, a general-purpose programming language, and OPAL, an operational research platform.
- Python is eating the world: How one developer's side project became the hottest programming language on the planet (TechRepublic PDF)
- Getting started with Python: A list of free resources
- Is Python's popularity peaking? Will the programming language's limitations end its rapid rise?
- A beginner's guide to Python: Books, tutorials, videos, use cases, and developers' favorite tools
- Microsoft: We want you to learn Python programming language for free
- Programming languages: Python overtakes Java on GitHub as Google Dart use soars
Roles on the team: Within the team, there are various roles. First, there's the machine learning expert with that mathematics background to support it. There are also roles for a data engineer, a person that would know Scala, Python, and databases such as noSQL and Hadoop. There are also a small number of "decision optimization experts" and visualization engineers. These visualization engineers, also known as data journalists, have a visual design background with some understanding of machine learning.
The screening process: The initial screen for the Data Science Elite Team is a coding challenge that candidates complete on their own. If a person passes that hurdle, there's a coding session via video conferencing with one of the more senior people on the team. Through an interview process, a candidate is screened for cultural fit and ability to be client-facing. "You can have expertise, but not be conducive to being in front of clients," said Dobrin.
Soft skills: IBM doesn't test for soft skills, specifically, but to be client-facing, a candidate has to have good communication, presence, and good behavior. IBM also wants to pose problems to see how a person thinks through a problem when an answer isn't known. It also helps if a person likes to travel since 75% of the job is on the road.