There’s a lot of former academics switching careers1 to the tech sector these days, and as an example of one such person, I was recently asked to write about the experiences of a former astronomer at Poplatek. So what is an astronomer2 doing as a software architect, anyhow? Well, let’s talk a bit about the cloud to start with.
Image 1. The Pillars of Creation – a famous photo of a star forming region in the Eagle Nebula taken by the Hubble Space Telescope.
There’s quite a lot of talk about moving legacy infrastructure to the cloud, but did you know that some of the oldest (and arguably most important) infrastructure around was deployed in a cloud?
I’m talking about the Sun of course. Like other stars in its weight class, the Sun began its life as a collapsing cloud of gas and dust in a fragment of a probably long-gone molecular cloud called the pre-solar nebula3 and, following a brief4 warm-up period of about a hundred thousand years, the Sun began running in (energy) production, boasting a current uptime of about 4.6 billion years.
While many would argue that nothing about the scale of services provided by the Sun could be classified as “micro”, it is still among the most compact and lean known implementations of this life-critical architecture; the Sun is actually classified as a yellow dwarf star. This also has not discouraged several past, ongoing and planned attempts at providing an even more compact solution to some of the services provided by the Sun which, if successful, will disrupt many of the industries the Sun currently dominates in.
Regardless of even smaller-scale competitors, the sheer scale of the services provided by the Sun still have an irreplaceable part to play at maintaining good service quality in the single availability zone currently occupied by effectively all known active commercial ventures in existence. It can only be hoped that we manage to achieve increased redundancy by spreading to multiple availability zones before the Sun enters its end of life phase in about 5 billion years which requires a full rebuild in a new cloud environment.
But wait: aside from delivering a terrible collection of puns5, I haven’t actually explained at all what an astronomer is doing working with internetty cloud architecture, or how an astronomer could be even halfway competent at it. So what’s up?
What I’ve been trying to get to in a very roundabout way is that the process of training a PhD in astronomy or probably any natural science6 requires of or equips a person with a few of the more critical (in my opinion) skills that are also needed in software architecture. To name a few:
- Looking at big unfixable problems, and breaking them apart into smaller more fixable problems
- Thinking about complex interconnected systems
- Curiosity, and the willingness to keep learning new things
- Some sort of programming ability
- Writing stuff down
Of course some extra work is required to properly qualify for such a career change. Namely stuff like:
- A basic understanding of how computer networks work
- Writing programs that are actually a bit maintainable
- Knowing that version control is a thing7
Lucky for me, I already had an interest in computers and programming aside from astronomy, so I’ve slowly learned how to behave over the years.
At Poplatek this has lead me to spinning up some machine learning models because I know statistics and data science. I’m also working on the backend of Poplatek’s brand new IoT platform because I know how to deal with large amounts of sensor data. Furthermore, I’m generally getting my hands dirty with AWS cloud architecture because that’s what I want to learn.
It’s been an interesting journey of continuous discovery, of exploring strange new stacks, to seek out new data and algorithms. To boldly code what no one has coded before.
Do you need expertise in IoT or cloud architecture? Aleksi and our other experienced specialists can help you. Contact us!
- The why of it is a different topic, but many people doing the switch often link to this article at some point.
- Astrochemist, to be precise. Here’s my thesis in case you need something to help you with falling asleep.
- Astronomers are not very good at naming things.
- It’s brief on a cosmological timescale.
- Which I love in this context, because I actually am an expert in a certain aspect of space clouds.
- And maybe the humanities? Sorry, I don’t know much about how those fields work.
- Fun exercise: Try to get your hands on a piece of code used to analyse data for some important astronomy paper. You’ll probably have to email the author to get it, and you’ll probably receive a single gigantic IDL script that’s called something like “analysis_ver6_paperversion_final4.pro”. Bonus points if you can actually get it to run.