What jobs do they do?
Most important thing to know
All the below hinges on your PhD research/work being relevant and allowing you to add value to a company.
Being super awesome at something that no one makes money off doesn't help you. Spending your 4-5 years of your PhD being a Teaching Assistant and writing a dissertation on something you don't want to actually do longer term is a really, really terrible idea.
I can't emphasize this enough. Too many people go to school for a PhD and then don't get any translatable or marketable skills and effectively waste their time.
Many research positions actively recruit PhD graduates. A non-PhD may not be able to apply for many of those positions, where demonstrated research and specialized knowledge is desired. You might even need post-doctoral work too.
Now if you aren't lucky/good enough to get into those positions your PhD can actually hurt you for "entry level" positions. You will be competing against people who generally are more qualified since they are more recently familiar with the technology. They probably have done internships in industry much more recently. They will be cheaper, or at least, have lower expectations.
Keep in mind there are often many people competing for a relatively small number of actual jobs. If you go to graduate school to get a PhD to teach, you might be in for a rude awakening as you find out there are considerably more graduates than positions.
Sometimes for startup companies the PhD can provide value. Assuming your PhD provided meaningful research/useful skills, the PhD process requires a lot of work which can be beneficial for employees in startups - whether pitching your ideas, writing proposals for funding, mentoring junior students, connections with others doing similar work, or just the general independence and self-motivation required. All these are really beneficial in different startup environments for those with the right background.
This isn't for everyone and isn't every startup.
Do PhD holders apply to junior software positions when they finish?
The only time you should do this is if your entire PhD was a waste of time. Unfortunately, this is often the case so sometimes it is true.
Your academic career should be connecting you with others in the field you want to go into (see below). By attending conferences, publishing papers, and reading research you will know who/where the relevant research is in the world. Your advisor matters. Your university matters. Because you know what you want to do longer term, by the time you graduate you should know and be intimately familiar with who else does your research.
This translates into people who 1) know the domain you are learning and 2) might actually hire you for your PhD knowledge. So when you graduate you already know who is looking for your specific knowledge.
These are reasons why it's important to know what you want to do or at least have a vague idea.
Otherwise, you're back at the starting point you were graduating before - except now you have negatives and no positives.
The bigger question
If you want to do a PhD in a technical field you should know:
- What type of research you want to do and why. If you don't have something specific you want to do, your PhD probably will be a waste. This means you do research on something you want to work with after graduation. Spending 4 years and lots of lost earning potential to do a project you never use is a waste of your life but something many PhD students do all the time.
- What benefit doing a PhD offers over working (or a masters). In most cases you are better off from a pay/career perspective doing a masters. If you can't answer this either talk with faculty, industry related to what you want to do, or your school advising/career departments.
And if this all sounds really complicated and hard? Well, you're damn right it is. A PhD should never be a "I want to learn more and then get paid more eventually." If that's your goal you should either do a masters or work fulltime for a while.