Today we’ll talk a little bit about what a pipeline is and how to optimize it.
Pipeline is a term from computer science that refers to blocks that process data, which are usually connected in series.
Imagine that you get data on the input, then you process it, and the output goes to the input of the next block. It’s a bit like an assembly line… You’re assembling a product and at each station a worker is doing one activity. He/she gets a part of the product on the input, adds something here and there, and passes the product further on.
Pipelines are common and used everywhere, mostly to process large amounts of data. Of course, also in machine learning.
Anyway, now on to why I’m writing about this. Big companies like Microsoft and Google try to solve problems efficiently, without employing emotions. Their hiring process is in principle a human pipeline.
Candidates are the input data — people, admittedly, but to the company they represent just data and numbers. The first block, for example, is an online coding excercise that filters out a large portion of the candidates. We then move on to the next blocks, each representing one round of interviews. In this case, the blocks can be in parallel because at the time of the interview the result from the previous one is not yet known. And at the end the successful candidate gets an offer.
Automation in practice. In a way, I quite like this approach. You’ll encounter also tweaks such as each interview is with people from different development centers from different countries, so that there is minimal bias affecting their decision.
I’m in a different pipeline now, I’m on a dating site. The input is my profile, so a photo and a description. When you look at it that way, you can look at it as an optimization problem.
How do I get as many women to write me back as possible and potentially have a better chance of finding one, ideally in a reasonable amount of time? By adjusting the profile.
By tailoring your profile to stand out among others, you’ve been able to ensure a higher percentage of success. You do the same in the next steps. For example, the first message to a girl on a dating site… Someone will say: “Hey, do you want to text?” And I’ll focus on a specific thing in the profile that I’m interested in and ask her a question. There’s already a dialogue between us.
Let’s go back to computer science. I’ve mentioned this before, but a lot of things in computer science are inspired by the real world, and vice versa. And sometimes it’s also nice to look at things as problems that we want to solve and optimize. At least because we have something to think about again 🙂