However, I know that I do not want to do Ops work professionally. I kinda just fell into it right out of college and chose it because I saw the potential.
However, I feel like that knowledge has gone to complete waste and wonder if I could leverage it in some unique ways.
What would you recommend that I do?
It can range from developer, to pure support engineer, to a jack of all trades, to yaml engineer, to CI/CD maintainer, to Google defined SRE.
Figure out what you want to do, and what you dont want to do, filter job postings and ask questions early in the process.
I'm a "Lead DevOps Engineer", and what that means at my company is that I lead a team doing all of our infrastructure automation(terraform, python), CI pipeline creation and maintenance, update and maintain all the non-prod environments (k8s, docker compose, swarm), infrastructure architecture on all major new projects (since we'll be building the automation to bring those projects up and maintain them), we also get everything else that seems hard to other teams. We build the tools and automation that we use, as well as the tools other teams use. This is important because we provide a common interface to all our teams, even though we work across AWS, Azure, OCI, and on-prem. The developers dont care which cloud their stack is on, that is abstracted away from them.
What we (DevOps) emphatically do not do is on-call. We have a "SRE" team for that (not SRE in any sense of what you think, they deploy new app versions to prod and manage datadog, and manage on-call, they are incapable of writing even basic python).
This sounds like one of the most screwed up environments I've heard of in a while!
Which is even more annoying because every sre job is some "everyday is different" bs.
Starting with firing 90% of them once all processes are automated away and they can no longer justify headcount(3x the headcount of DevOps). It will be a battle, that VP has nothing but SRE under them and the only reason they werent fired years ago is because of their connections with the CEO.
I can really recommend this path.
It'd be helpful to know why you don't want to do Ops work. There are many roles that fall outside of DevOps that have the same downsides (frequent oncall, unpredictable workloads, further from the product, CEO doesn't understand what you do, etc). There are also roles that fall under DevOps that don't have some of those downsides (infrastructure architect-type roles).
For example, there are people in the comments recommending security, which is great for job security, but generally bad if you want to do interesting work. Usually developers do most of the actually interesting security work while security teams check boxes for SOCII audits unless you're working for a security product, which can be pretty competitive.
Now, I feel like i have such unique knowledge, application of k8s at such a large scale, and I no longer want to use professionally. Makes me feel like my career till this point, has basically gone to waste.
I have done oncall, CI/CD, Kubernetes, dev tooling, terraform and some other devops stuff.
Actual in-the-trenches experience of distributed systems at scale is fairly rare. Anything that helped you acquire that while being paid is likely not wasted!
I’m assuming by this you mean non-technical people.
It feels like if you can handle the chaos, an early-stage startup is the way to go for you. A single developer can significantly contribute to product discussions and infra platform decisions simultaneously.
Perhaps also look into positions outside of traditional tech? Med-tech, green-tech, etc are all in desperate need of infra experience, but tend to not have siloed roles like tech-tech has.
Often Developers see security as a distraction from building features, so working with companies to ensure they have strong security is a good career.
I worked with a bank building a "secure cloud" on top of a public cloud. That is, our team got a long list of requirements, like "storage buckets encrypted at rest", we'd watch public cloud changes, then alert or fix the issue. The bank got to use the public cloud, but also had a much finer-grained control of a lot of security concerns.
Personally, I find it alarming and problematic (though understandable and not surprising, obviously) that so many industry (services, jobs, etc.) are dependent on the big three public clouds (G, A, and M).
- Those owners (and their tenants) can decide to boot customers on whatever (illogical, political) whim; I think this is not purely fantasy.
- Despite whatever portability may exist due to the containerization and microservice standards, those owners essentially practice the "capturing" of customers (and workers) through pretty complex sets of knowledge for running, deploying, and maintaining the services. (Always be upgrading, whether for features, security, API/tool/ecosystem/business updates, etc.).
Sure, there are many other hosts. But I think there is (maybe) something where one could say:
1) Okay, it is (stupid and uncomfortable) to depend on G, or A, or M, given all that they do, even if they abstract away a lot of details the industry has come together to solve.
2) The solution is (stabilize, document, and publish/educate) regarding a set of tools/practices for being truly "infra agnostic"--to build and run the same microservices, but on any given set of hosts that one can own/rent/etc.
For background, and some of the above motivations aside, as someone who worked in a big company that owned its infra and did not use the big 3, and seeing all the jobs requiring experience with (one of) the big 3 public clouds, even though studying up would be doable, it seems like a fool's errand to me and a huge turn-off to learn even one much less all of the big 3 public clouds' ins and outs, because I know they will just change and I know for sure they are not thinking companies that will do the right thing.
Anything you could comment on regarding whether this is or is not a real problem (or related to one) that you might be uniquely able to solve by leveraging your experience would be very interesting to hear.
You'd be amazed (by the sound of it) at how little changes, at least in AWS.
I'm not convinced that people "forgetting" how to run hardware and network gear is a real problem. Most people never knew to start off with.
If it is a real problem, a market solution will emerge that replicates the API-driven provisioning experience of public cloud on hardware. Several companies are working on this, and we'll see if they are successful or not.
My low leverage answer was to make awesome youtube video tutorials.
Thank you so much for your response!
I’m not really in this area, but something like Google SRE handbook might get you excited about it.
I am also shipping my ideas for the first time ever. Super excited about it.
- learn server side language(s) & become backend dev
- get an IT sales job