A Haystack Full of Needles

Imagine you were searching for the world’s absolute best software engineer. The second-best isn’t good enough; your client insists on the very best. Other than Googling “linus torvalds contact information”, how do you do it? Presumably, you look for a lot of direct signals: looking at the commit history of the repositories for major open source projects like React or TensorFlow, for example. You probably also look at the staff lists of the major tech companies. And in doing this, you’ve generated a huge pile of engineers. Now what? You could give each of these engineers an incredibly difficult coding exercise through HackerRank or something. But presumably they’ll almost all score very well; that won’t provide any particularly interesting information.

We don’t (usually) get the brief that we need to recruit the world’s number one software engineer. But the problems associated with it are actually pretty similar to a normal search. The problem, for most technologies, isn’t generating the list, but winnowing it down, or generating a smaller list in the first place. We do some of this the way that we looked for the best engineer in the world above – looking through public commits and previous places of employment. But there are also indirect signals that can help.

  1. How familiar are you with the educational structure in the area you’re hiring? There’s more to know than just the best colleges in the area. If you’re hiring in France, for example, you should probably be familiar with the secondary-level system.

  2. Are there standardized tests that most of your candidate pool might have taken, even if those tests are not directly related to engineering? In general, if your goal is shrinking the pile, even signals like a high LSAT or MCAT score is a good sign – that gives you reasonable evidence that the person is strong at abstract reasoning.

  3. Outside of structured educational or career attainment, is there something that the candidates do excellently? We’ve hired people at partially on their strength of – real example! – writing crossword puzzles.

None of these are guarantees of quality on their own. And in fact, the best software engineer may not fill any of them – Torvalds didn’t go to MIT or Stanford. But our goal shouldn’t actually be to find the best software engineer. It’s to generate an excellent software engineer, following a process that’s repeatable and quick. And to do that, you’re not looking for a needle in a haystack. Instead, you’re looking to avoid ever generating too big a haystack in the first place.

Want to talk about the right hiring strategy for your organization? Book a call with me here.

Previous
Previous

Stop Not Building Your Product

Next
Next

Engineering, Coding, and AI