Technical knowledge isn’t a uniform thing. Far from it! Instead, it consists of different layers of knowledge. For hiring teams, a key question is: how do measure these different layers of knowledge when assessing candidates for your technical roles? In this article you will learn about:
- The three layers of software development knowledge
- How to measure each layer of knowledge
- When to automate and when not to automate knowledge assessment
Technical knowledge can be thought of in terms of three different layers: the core layer, the language layer, and the framework layer (see our last blog post for a more in-depth discussion of each of these!). The core layer refers to engineers’ basic programming skills, while the language layer refers to their in-depth knowledge of a specific programming language. The framework layer builds on these by addressing engineers’ abilities to build a working application in a specific framework.
Recognizing that technical knowledge consists of different layers is an important first step for your hiring team. But this still leaves an important question unanswered: how do you measure each of these layers of knowledge when assessing your candidates’ technical skills? ????????
The answer is in taking different approaches to measuring core knowledge and measuring outer layers of knowledge. It also requires you to know when and when not to automate your assessments depending on the knowledge layer you’re measuring.
Measuring Core Knowledge
Core knowledge is the common programming knowledge shared among engineers regardless of the programming language they use. To use a sports analogy: the core is like running. Both soccer players and basketball players, for example, use running as a fundamental skill. Learning how to score a goal or strategize on the court, on the other hand, are more specialized layers of sports knowledge.
When you measure your candidates’ core knowledge, you want to see their ability to translate ideas into code. One example of measuring core knowledge, says CodeSignal CEO Tigran Sloyan, is asking a candidate to perform a basic coding task like merging strings.
Sloyan also recommends that companies automate their assessment of candidates’ core knowledge to recruit more efficiently.
Tools like CodeSignal make it easy for you to automate the assessment of core coding skills. “Ideally, you’re looking for leverage in your ability to scale recruiting and to measure ability,” says Sloyan. “When you measure the core, it’s significantly easier to automate because there’s not too many variations.”
Measuring Outer Layers of Knowledge
Measuring your technical candidates’ outer layers of knowledge, language layer and the framework layer, is a bit more complex. But depending on the role and level you’re hiring for, you may be able to automate your assessment of outer layers of knowledge – particularly the language layer, says Sloyan.
At the framework layer, measuring candidates’ knowledge requires a more human-driven process. Companies will still benefit from using a live interviewer to think and talk through decision-making processes with the candidate. A team hiring a React developer, for example, will want to know how the candidate thinks about the design, scalability, and latency aspects of their application.
Rather than fully automate the assessment of the framework layer, Sloyan advocates for using an “AI-assisted” approach. This is similar to the AI assistance features in many cars today, which alert you to a car in your blind spot or an object you’ll want to avoid.
Sloyan explains, “When it comes to measuring the outer layers, today the best approach is to do an AI-assisted version and not a full AI version.” An AI-assisted approach offers the best of both worlds for your hiring team: the efficiency and consistency that comes with automating assessments, and the ability to customize your assessment through a human-driven coding interview.
When thinking about how to measure different layers of knowledge with your technical candidates, keep two key principles in mind: 1) use different assessments to measure different layers of knowledge, and 2) think carefully about when and when not to automate your assessments.
According to Sloyan, “it’s only a matter of time” before all layers of technical knowledge assessment can be automated. In the meantime, he says, “the best thing to do is to augment some of your interview processes with AI-assisted assessment tools.”
Want to learn more about how you can build a winning organization through data-driven recruiting? Visit CodeSignal to find out how you can measure technical skills effectively and objectively with its automated assessment and live interview solutions.