The single algo program is a proxy for "can this candidate think through things logically and come up with a good design, and explain it". If you can do it, th

The single algo program is a proxy for "can this candidate think through things logically and come up with a good design, and explain it". If you can do it, there is hope you can learn docker, k8s etc. However you can be "good" at docker, k8s from courses without being a competent programmer (but a good systems admin). Also a company has to deal with 1000's of applications and unfortunately they need early filtering stages, they don't have the resources to look at that many Githubs. Let's say it takes on average 30 minutes per Github to evaluate - that 500 hours per 1000 candidates, of which that time is coming from people who should be doing Engineering tasks. Even if they can afford it, it will slow down the process too much. And then how to evaluate Githubs? - did they just fork something? Was it their code? Did they get help? Etc. Github is maybe OK for a later stage interview where the investment is worth it to discuss the projects with the candidate in an interview. It could be a nice alternative to a coding test. But you need an interview to get a feeling for if they really wrote the code. My advice - be a good dolphin, and learn to jump through this hoop. It's like the driving test where you need to prove you can "reverse around a corner" - something I have never done while actually driving.