IT worker shortage: age discrimination and unrealistic standards.

You may have heard about the difficulties of some tech companies, specially when it comes to hiring. Well, is my duty to tell you, companies are either not being honest or not being realistic, taking into account they are built to make money, I’d like to think the second option is not that likely.

Let’s start with me, I’m a 25 old graduated, with a master’s degree and just shy of 3 years of work experience. The position that describe me the most is modern DevOps, I know how to do the Kubernetes, absolutely love docker, can code in Python and Go with 2.5+ and 1+ years of experience respectively and currently learning Terraform. Not bad, isn’t it? No, is not enough for your average fintech.

This ‘DevOps’ role, (not to be mistaken with ‘modern SysAdmin’), is highly sought after in the startup world, at least is what most recruiters say, therefore, Why can’t I (and many others) land a job in one of these companies?

The answer is the same as always, insanely and unrealistic job requirements. You may be familiar with Tiangolo’s job posting anecdote, sadly is just the tip of the iceberg. Back in 2019 I remember seeing job offers with 5+ years of Kubernetes experience, having into account K8s was release in 2014, that narrows the possibilities to just ex-Google engineers.

Nowadays your average startup looks for profiles with 5+ years of professional experience in extremely young technologies. And let’s put it clear, those profiles are not for hire, companies will need to decide if they want the technologies or the years of experience. Obviously they’ll choose the second one and hope that engineers that have been doing the same thing for 5+ years will change the way they solve problems to adapt new technologies.

Some companies have begun realize having a staff of full seniors is unrealistic and opted for a different approach, 60% seniors, 30% mediors, 10% juniors. That’s much, better, but also unrealistic. Let’s go with the numbers! based on the estimations of Robert C. Martin, the number of Software engineers duplicates every 5 years. Flip that out and you’ll get that in any given moment half the software engineers have less than 5 years of experience.

Let’s analyze the average job offer

This job offer corresponds to a Senior SRE at a mid-to-large (500-1000 employees) company founded in France 2014 offered in Berlin. Let’s have a closer look. Line 1, requires 4–5 years of experience, as said before 50% of the potential hiring force had been filtered, but there is more to it, ‘in a similar role’. Let’s assume there are 3 big roles, Software engineers (Back, Front and Fullstack developers go here), Sys Admins (Automation Engineers, DevOps, …) and Data Scientist (AI experts, Big Data developers, …), assuming an even distribution (not really, Software engineers are overrepresented), that filters another 33% over the previously 50%, i.e. Only 15% of the active workforce at best and being generous can past that very first line.

A quick search of position of ‘Senior SRE’ and ‘Junior SRE’ jobs in Germany, has devastating results: 2 Junior offers vs 58 Senior offers. That makes even more difficult that apparently easy sentence ‘in a similar role’. I don’t know if how will that translate into exclusion, therefore, I’ll be extra generous and not take it into account.

Lines 2 and 3 are no big deal for the role, let’s jump to row 4, ‘Solid experience With Docker / Kubernetes / Terraform / Datadog’. The release dates of those technologies are 2013 / 2014 / 2014 and 2010 respectively, given that requirement, ‘4 -5 years of experience in a similar role’ that pretty much filters anyone but early adopters. But, let’s be fair, there are high chances that most of these profiles are into those technologies. Of that 15% of available workforce, more than 80% should be familiar with the K8s ecosystem. Down to 12% of the workforce.

I’ll comment 5 later, let’s go directly to lines 6 and 7. Chaos suite and Prometheus are particular tools for general problems. The case of Prometheus, is stupid, there are gazillions of monitoring and visualization stacks, using that one instead of TICK or Datadog is just chance, same goes with Chaos suite. If that line is serious, if the important thing to know is Prometheus, not monitoring and visualization, the percentage goes from 12 to less than 1%.

Languages, my fave. Knowing bash is a must for this kind of roles. On the other hand, asking for Python, is not bad, but any scripting language will do the trick. As before, I’m not gonna be harsh, knowing TypeScript or Perl is close enough to knowing Python (for this case), depending on the day you may or may not be hired for that. At last we reach Go. Go is the hardest language ever, it requires you to know Software Engineering, it requires you to know Data Structures and Algorithms to work it at is full potential and that’s why I love it, Go is not an scripting language. On top of that, less than 9% of the developers know go, according to Stack Overflow statistics. My dad always said ‘leave the best for last’, so Go looks important. For sure the profile is quite straightforward so won’t be unreasonable to think that Go developers are overrepresented in these kind of roles, lets suppose 200% more Go representation.

Results: This ad targets 0.2% of the workforce, and that’s being extremely generous.

And the sad note is that this is the industry standard.

I’m in that 0.2% of that workforce, what’s left for me?

I’d like to say that everything, but let’s read carefully line 5. A senior engineer, a Kubernetes early adopter, a Go machine, an SRE before SRE was a thing will for sure have the privilege of… being available on weekends. The job market is broken.

Ageism, experience and expectations

Summarizing up everything, you need young people, if you want to build your startup from young technologies, otherwise, asking for 5+ years of experience with Go and Kubernetes limits the offer to ex-Google engineers and 4 lucky early adopters.

On the other hand there are pools of junior developers that have learn how to code with Go, that had known how to dockerize their code since day 1, that understand K8s ecosystem as the game changer it is that are not given the smallest chance.

In the end an engineer that writes Go as it was Java, that doesn’t get the difference between musl and glibc in docker alpine and that doesn’t understand how the state of a VCS repository can reflect the state of a company’s infrastructure will still be more attractive than a junior.

Is there any solution?

Nope, there aren’t.

FAANG companies where built by juniors making there own rules, startups nowadays try to emulate that with seniors, doing what they do, is not going to work. Rules can only be broken once, and these companies are building planes, not inventing them.

Having said that, if you are in one of these jobs, ask for a raise, now, and if you aren’t start applying.