I can't specialize, how to differentiate myself?

Psychological skills will help.

I have been working in unfamiliar territory for so long that it's become one of my strengths. The first skill to develop in this domain is that of acceptance. Accept the unfamiliarity. Accept that you are going to do the work. Accept that the process is going to painful at times, and realize it isn't going to be painful all the way. After you accept, consider how much further you may go if you actually embrace your work. You're going to have to do it, so why not give your best? This may likely require you to put more into the process but the accomplishment is how you reward yourself.

You can become a generalist with specializations. You don't have to only be one or the either.

Highly recommend to read this book:

Range: Why Generalists Triumph in a Specialized World

The author went on lots of Podcast shows, you can start from there before you spend any $.

There is nothing wrong with being a generalist! You could just focus on cloud technologies and it would occupy a lifetime. Deep dive into linux api, systems programming, risc-v, fpgas, etc

MeganWachs - Keynote RISC-V and FPGAs: Open Source Hardware Hacking

https://www.youtube.com/watch?v=vCG5_nxm2G4

And just keep diving. Get low level. I think next gen cloud architectures, gpu, tpu. Next gen lidar, radar. Playing with 4D point clouds in VR. That's about as hot as it gets right now ;)

DigiKey & AdaFruit "All the IoT"

https://www.digikey.com/en/resources/iot-resource-center/vid...

You absolutely can specialize, you can even describe your current role in a way that sounds specialized - "Build/Release Engineer specializing in CI/CD and cloud native technologies."

As far as specialties inside DevOps - security, OCI, schedulers are general areas that have specialist roles available. Another option is to specialize in a tool or tooling ecosystem relevant to DevOps. For example, you could decide to learn AWS service offerings inside and out or specialize in webtech on bare metal and land a role at a company that has the scale to demand optimal bare metal deployments.

There's value in marketing yourself as someone who has a variety of domain expertise. Having a wide portfolio of projects is attractive for companies with uncertain needs, or in new domains.

I'm in biotech, and being the "software guy" on my team has left me in a similar situation, and I asked the same questions for how to market myself to Ph.D. programs, which are all about specialization.

What kind of problems do you like to solve? Figure that one out and then learn the tools/skills required to solve those problems.

Something like "MLOps" maybe? https://towardsdatascience.com/the-rise-of-the-term-mlops-3b...

Or, if you have time, contribute to some Open Source tool you use or like. It can help with differentiating.

You don't provide any context as to what form your struggle takes with each new project. So it is hard to give general advice. However ...

In doing the many random things, have you noticed any that you enjoy far more than others? If so, then it might give you an ideas as to what area to focus your learning and development upon.

Do you make any of these random things easier to deal with the next time around? If so that's your specialty, otherwise you could start thinking about ways to do less to get more done whether it's with scripts/tools, other software, documentation, interacting with others, etc.

A DevOps Engineer is a pretty specialised job in the grand scheme of things.

Money and accounting.

Not to deep, just understand the difference between opex, capex, and what implies to your job.

This will make you able to speak with the business inthwir own term.

You can specialize in any area you wish to. Just find what you like and put your interest in it during your slow moments or after work hours.

What do you want to specialize in?