Hello,
TL:DR Atmospheric scientist with knowledge of cli, python want change to a SRE job
Let me explain my background, I am M.Sc in atmospheric science with a few publications in the field, during my studies I work most of my time working with cli tools so I am confident with shell and cli tools, pipelines, stdout, stderr, tmux, etc. I worked using numerical model that you need to compile, this teach me tools like makefiles, modules and cronjobs to did it operational. I have experience with python and other scientific languages like R, Matlab. During my free time took some course of docker. Even I set up a Nginx webpage that I leave to die for lack of time or setup a raspberry pi to download “linux isos” using docker.
I really enjoy the automation of process, and help other colleagues to setup and install the environment to work.
I know my lack of networking, monitoring, and I’m not sure if my self taught skills (science standards) are comparable with a CS worker.
I want to learn or get the certification needed to get a SRE job to mid 2024.
Any advice, course or certication to help me to get in the road?
I think you have an interesting background and potentially interesting technical skills, and I could totally see you catching on with someone and having a fantastic career. I could also see why it would be a weird or awkward fit, that you might be totally overwhelmed, and possibly even hate it. Let me qualify my answer(s) and see if that helps at all.
I feel like at its heart, being a DevOps is just being passionate about tinkering and technology. The best DevOps Engineers I know love nothing more than to nerd out about…well, all kinds of stuff. From K8s to Linux distros to build tools to code. DevOps is a practice, not a skill set - and that’s reflected in the fact that there’s no ‘base’ skill set for DevOps Engineers. I’ve known developers, sysadmins, even help desk type folks that found their way into the field and were successful. It just depends.
It kind of feels like you have the heart of a tinkerer, and the fact that you have a MS in a hard science suggests that you have the brainpower to hack it - maybe literally. :)
That said - what would worry me if I were considering hiring you is that you don’t really have any exposure to Software Development Lifecycle concepts. Maybe I’m too stupid to understand all the acronyms above, but in my (limited) experience, having a good handle on SDLC is sort of the beating heart of DevOps - at least in part because being able to have the infrastructure ready to mate up with the code at the right time and right place is like, 80% of my gig. Too early is a security vulnerability (potentially), too late and the dev team misses all their sprint targets. You don’t have to write code, exactly (although I wish I wrote more), but you have to be able to ‘follow along’ with the dev team. Especially when you’re troubleshooting.
For SRE particularly - you have a lot of nice sysadmin-y type background skills, but particularly understanding design patterns and telemetry would be the thing I’d be most nervous about for you. Scalability as well - although that’s hard for almost everybody. But for an SRE to improve reliability, you have to be able to really hone in on what’s breaking - and once you’ve gotten the big pieces sorted, being able to understand resource usage, and all of that points towards good instrumentation (and good instrumentation practices).
I joke that reading logs is my superpower - both because my devs, bless them, don’t do it, and also because if we’ve done a good job building the application, build/deploy pipelines, and infrastructure, your alerts and instrumentation will tell you exactly where any pain points are happening, and make it a lot easier to figure out where and how to focus your efforts moving forward.
So, after that wall of text - I’d point you towards the cloud. AWS is the largest/most widely known, but arguably kind of opinionated in terms of implementation. Still, AWS Solutions Architect is a pretty good ‘gold standard’ type certification. If you’re more familiar with GCP or Azure, do the ‘associate’ level certs there.
Another obvious thing that I didn’t see in your background - VCS. Git gud, as it were. I’m a big fan of hanging pretty much all your personal projects on GitHub. Mine is atrocious since I got hired, but before that I had a full year straight of commits. Sometimes it was impressive stuff, most of the time it was just messing around with code - but all the companies that gave me an offer letter mentioned it. Ymmv.
Finally - you might expand your search a little wider (SysOps instead of SRE off the bat? DevOps as well? Maybe going straight stick software dev, with your background, at a company where your science background would be a real value add is something to look at) and also be prepared to ‘take a step back’ if you do jump. I’d definitely hire you to see how things go, but I’d want you to come in as a Junior, and based on what you wrote above, that’s probably a bit of a paycut for you.
TL;DR - Do cloud certs, practice on GitHub so employers can see what you’re working on, consider SysOps/DevOps as well as SRE.
Best of luck to you!
Thank you so much, you did a fantastic job finding my weakness.
Yes, I’m aware of this, I need to take some time to learn it and get familiarize it.
Also true, hope some of these topics have an overview in a cloud cert.
I have similar experience with my colleagues when I was able to compile climate models, the difference was I spent some time reading the logs and look what’s dependencies are missing or the path is wrong
Thank you I’ll take this one
I have some experience using git and svn, but really never work in collaboration with others, in my current work we used but only git without external service. Just to keep track of the personal work.
I need to use it more, I only use it for “more” important projects, Now I think is bad approach
For sure, I’ll try a wider search, I always have troubles to find a job offer where I could met the requirements, thank you for suggest 2 jobs :)
I have a “software engineer” job in a research institution, is only the title because I’m a research assistant most of the time with some dev time. The problem is there is no grow and the founding is through a international project, so the time is fixed.
My pay is not so high, is a EU salary in a semi-public institution, tho the pay is lower that the equivalent in the industry, but I’m above the average of at county level, I’ll consider a paycut at least I could still pay the rent and past time with my family.
Same to you, you are very kind with extranger a little lost about his future, I’m appreciate a lot the effort of you reply.
No this is great in and of itself - what I would tell you is to treat your github projects, even if you’re the sole contributor, like you’re working on a team. Checkout a feature branch, complete your code, then PR it into main/master/release/whatever, even if you’re the one doing the code review for yourself. Even if you don’t get to experience other devs (inevitably borking what you’re working on) in the codebase at the same time, it’ll give you a better idea of what the workflow should look like.
This all makes the SRE part more understandable and more within reach. I wouldn’t lead with, “I don’t have any dev experience”; I would lead with “I’ve been a software engineer for x years, specializing in atmospheric modeling.” Whoever is interviewing you will probably dig and figure out that you were a solo developer, but…you were still a software dev, and the first job in this industry is way harder than the next couple. Lean into that - you have the job title, you have the resume, you’re looking to take ‘the next step’ into SRE/DevOps, because as a solo-dev, you had to handle all that stuff yourself, and you figured out that you liked it and were good at it.
We’ve all been the new guy trying to break into the field - pay it forward after you land that first SRE gig.