DevOps - you'll quickly run into requirements to write parallel work. Not using something like fabric is a waste of your employer's money hiring you. You'll quickly run into requirements to use ansible/terraform/other. Same applies here; surrounding tools with subprocess and parsing stdout/stderr is a waste of your employer's money. Use appropriate packages, instead.
DataScience - all the cool toys, from simple stuff like pandas, or more focused ones like working with spark and the plethora of big data libraries require indepth knowledge of both the library and the underlying datastore. Combined, you make the most of your time and hardware. Not doing so, and only writing something that works on a small scale is a common costly mistake.
Security - python is just a helper tool. Your requirements, depending on position, will be focal knowledge about a domain and its set of tools. From highly skilled domains like research onward to threat hunting, PT and to the entry level stuff like NOC.
You're better off enjoying the little wins in life and dealing with the less pleasant moments. That'll give you a nice career in anything you choose.