It is important for AI/ML organizations at enterprises to consider opportunities for professional development of data scientists. In addition to coaching and mentorship, we have observed that exposure to a wide range of problem types and product areas is a critical requirement for the rapid professional development of scientists early in their careers.
Rotation of data science projects is key to give scientists an exposure to a wide range of problems and to build out their experience with different problem formulations, solution architectures, frameworks, and algorithms.
To maximize a data science team’s potential, consider regularly rotating scientists to new projects and products. While there are exceptions, we typically recommend that data scientists be moved to different projects every three to six months. Rotations are often a win-win – facilitating professional development while also bringing to bear fresh perspectives on projects and problems.
We also recommend a mixture of project work and product work as part of rotations. We find that data scientists may develop specific, useful generic or reusable artifacts while working on specific problems. We want to give the data scientist who identifies and develops the first version of an artifact – for example, a novel unsupervised anomaly detection pipeline – an opportunity to productize that service so that other data scientists can use it in their work. We therefore explicitly recommend product rotations for data scientists for more complex work, as depicted in the following figure.
Figure 46 Typical data scientist rotation across projects and products
We also recommend thinking carefully about the advancement and growth of data scientists across several dimensions, including core problem-solving skills, AI/ML skills, coding skills, leadership skills, and communication skills.
The upshot is that AI/ML is a human capital game. Recruiting, training, organizing, and developing a bench of talented AI/ML practitioners is a critical success factor for enterprises seeking to transform their business operations using AI.
Just as organizations must invest in building internal expertise throughout every functional area of the business from finance, marketing, and sales to research, manufacturing, and logistics, a strong data science team is critical to succeeding in the digital era. Organizations that excel at AI/ML, particularly those that take an early lead to build out AI//ML capabilities, will reap significant, sustained competitive advantages. Those that fall behind in AI/ML will fare less well.
We have offered here some ideas and concepts that may be useful to enterprise business leaders seeking to improve their organization’s AI/ML capabilities. Taken together with the rest of this reference, we hope we have presented an actionable and effective management guide to power successful AI/ML business transformations that capture business value at scale.