Home > Articles > How to train data scientists and engineers: a review of Peak

How to train data scientists and engineers: a review of Peak

Most companies with personal development budgets are wasting their money. If the goal is to help employees master valuable skills, then we are misallocating funds to books and conferences. Instead, we should look at what the research on skill development says so that we can make decisions informed by data.

Great people are hard to find, especially in software engineering and data science. The people who you want to hire probably work for your competitors and make more than you can afford to pay them. If instead of finding employees that are already great you could help employees become great, then the process of doing that would become a huge competitive advantage.

I recently read the book Peak by Anders Ericsson and Robert Pool. In it, the authors espouse the value of deliberate practice and offer research-backed insight into effective training practices. How might a company go about creating top-tier talent?

Before we talk about what works, let's think about what doesn't:

attending lectures, minicourses, and the like offers little or no feedback and little or no chance to try something new, make mistakes, correct the mistakes, and gradually develop a new skill

Unfortunately, many corporate training budgets are set up to send people to lectures and minicourses rather than building good training programs. The authors offer advice on how we might train doctors, though we could adapt their recommendations to the tech industry.

The first step is to find the experts. Ideally we could do this on a global scale, determining the greatest data scientists in the world. Unfortunately, even determining a methodology to find these people sounds prohibitively complex and difficult. Fortunately, we can settle for an approximation. Bringing an average engineer to the level of elite talent on the world stage would be incredible, don't get me wrong, but even getting that person to perform like the best engineer at the company would be a huge win.

Finding the top talent at your company may involve: asking various individuals who they hold in particularly high regard, examining performance review data, and/or determining which individuals have had the greatest positive impact on the business. If you're fortunate enough to have access to brilliant individuals outside your company, all the better!

Once we've found these experts, we move on to step two: understanding how they think about problems. Ericsson and Pool go so far as saying that the "main purpose of deliberate practice is to develop effective mental representations."

Having done this crucial work of understanding highly effective individuals, our third step is to build a "Top Gun" school. Modeled after an effective strategy for training fighter pilots, Peak's authors recommend creating training programs that simulate the real thing as well as possible while dramatically lowering the cost of failure. In our industry, this might mean identifying a few JIRA tickets representative of a team's work and having trainees work them under the watchful eye of high-quality instructors. These instructors should point out failures to their students and help the students to develop the thought processes (i.e. mental representations) of high performers.

Note that we're not developing coursework on design patterns, data warehouse design, deep learning, or a certain JavaScript framework. Instead, we focus on doing the work:

One of the implicit themes of the Top Gun approach to training, whether it is for shooting down enemy planes or interpreting mammograms [or developing a predicting widget manufacturing capacity], is the emphasis on doing. The bottom line is what you are able to do, not what you know

Building great teams isn't easy, but it is incredibly valuable. More than just spending $50 on eBooks, companies should create training programs that use data-driven insights into skill development to bring each team member to the level of our best performers. Great products come from great teams, and great teams are formed of great individuals. Great individuals are formed through deliberate practice.