In this special guest feature, Dale Kim, Senior Director, Product Marketing at Hazelcast, discusses how data scientists are a bit unique in that they have technical skills and deal heavily with data yet are not necessarily tightly integrated with the rest of the IT team. Dale is responsible for product and go-to-market strategy for the in-memory computing platform. His background includes technical and management roles at IT companies in areas such as relational databases, search, content management, NoSQL, Hadoop/Spark, and big data analytics. Dale holds an MBA from Santa Clara, and a BA in computer science from Berkeley.
With all the talk and action around machine learning, artificial intelligence, and data science in general, it is no surprise that the demand for data scientists continues to grow at a rapid pace. But the demand is growing much faster than the talent pool, creating a big hiring challenge for organizations looking to expand their data-driven initiatives. And, even if you are successful in building out your data science team, that does not mean your personnel problems are solved. After all, data scientists are a bit unique in that they have technical skills and deal heavily with data yet are not necessarily tightly integrated with the rest of the IT team. Many do not consider themselves to be coders, and do not naturally embed themselves into the IT culture and the software development lifecycle.
If today’s IT teams are expected to deliver business value on data, there needs to be a seamless way to integrate data scientists so that everyone is aligned as peers, rather than one group acting as a support center for another. So how does an organization create an interface between data scientists, data engineers, DevOps, site reliability engineers, analysts, and other IT folks?
Promote Collaboration
To overcome this challenge, the predominant approach seems to be to move from siloed data science and IT teams to integrated multi-role teams. We have often seen a centralized organizational structure in which the data science team is a cohesive unit that falls under a single leader. This is a useful model in that it allows the team to contribute to all business units, to collaborate closely within the team, and perhaps even to streamline recruiting.
However, it also reinforces the separation of data scientists from engineers. Possibly worse, it can reinforce a separation from the various business partners, as the data science team becomes a shared resource that is pulled in different directions. By opting for a more decentralized model where data scientists are dedicated to specific groups within the IT organization, it becomes easier to define common goals that help team members stay aligned. Collaboration is promoted within the integrated team, rather than having the data scientists in a centralized structure, pursuing a distinct set of mini projects in isolation.
Integration Fosters Relationships
Certainly, enabling this structure may be easier said than done. But this approach is not too different from engineering teams that already have members with diverse skill sets, including those who do not consider themselves to be coders. For example, technical writers who are in the process of creating documentation are an inherent part of the software development lifecycle. They do not wait to start their work until after the software is completely written. Rather, they work with engineering to understand the software’s capabilities and adjust along the way, as the feature list for a given release changes. Another example is the user experience experts who work with engineering from a design perspective. These team members also are not engineers per se but are deeply integrated into the entire development process.
A key part of successfully integrating data science with IT is having leaders who understand how to manage a team with diverse skill sets. Leading the team is not only about innovative technology and efficient development; it is about identifying all the ways that the software and data can benefit the organization.
The ideal leader has the technical background and the business-facing mindset to understand the business ramifications of the team’s work. That type of leader often communicates and collaborates with business partners to deliver on the business strategies. By incorporating data scientists as first-class members of the IT team, dedicated to supporting business initiatives, these leaders can help tear down the partitions that arise from the cultural differences of separated teams.
Sign up for the free insideBIGDATA newsletter.
Speak Your Mind