Are software engineers with PhD in demand in Academia?
tl;dr: Yes, some universities do value software engineers with a research background.
University College London (UCL), for example, has a Research Software Development Team (RSDT), which specialises in providing software engineering by programmers with a science-research background.
From the first of those links:
This is a team of professional software developers with particular expertise in designing, constructing and maintaining software for academic research.
Our goal is to enhance UCL’s capacity to produce high quality scientific software, from the simplest scripts to complex simulations running on state of the art high performance computers. We do this by collaborating with researchers who are creating their own software
How is it funded?
The work can be funded in one of several ways. It can be funded through university core funding. The work can be bid for as part of a research funding proposal or consultancy contract. It can be funded as part of a research grant for generic research software development.
What do they do?
The UCL RSDT team work alongside researchers on scientific projects. They co-author papers with them, transfer best practice & skills into the team, and hand over well-crafted, well-documented software (and its version-control history) for research.
The team do have research backgrounds themselves. They help select / develop the algorithms. When starting a new project, they read some of the background literature to the project - particularly that relating to the algorithms - to get enough grounding to enable them to ask meaningful questions of the researchers who they're working with.
And (touching on a comment by dgraziotin), the team host Software Carpentry Boot Camps for research staff and doctoral students. This gives the attendees an introduction to key concepts that they might otherwise not be exposed to, such as unit testing, version control, and working from the command-line.
Do many universities do this?
At the moment, this isn't an initiative at many universities that I'm aware of. However, given the success of UCL RSDT to date (the team is now expanding again), and the growing awareness of the concept of the Research Software Developer ("a new type of hero", as described by Neil P Chue Hong, Director of the Software Sustainability Institute), this is probably an idea whose time has come.
Disclosure
I am a little involved with the research software development effort linked to above.
Short answer: it looks ilke it, yes. I don't know where you are in the world but, when I've looked at academic jobs sites in the UK and USA, there have always been lots of ads looking for people to develop software for research teams – especially in bioinformatics but also in any area of science where computer modelling can be applied profitably. I've not looked closely at these ads but my impression is that, unlike the group discussed by EnergyNumbers, these are mostly fixed-term positions within a specific research group, tied to a specific grant. But the demand does seem to be there.