Conversely, why is it so hard to find good developers, people who really know something? For example, here's what we are looking for at Maxeler (http://www.maxeler.com):
Software Engineers: HPC Application Acceleration, Palo Alto, California
Application acceleration engineers work with client applications
to deliver order of magnitude speedups. You'll be involved
in every stage of accelerating applications, from analyzing
multi-million line code bases through developing new algorithms,
to implementing complete solutions running on FPGA, GPU, or
other technologies. We are in the early stages of growing and
building up the Palo Alto office, so you'll be part of a very
small team working in a start-up-like environment.
You will need a degree or equivalent professional qualification
in Computer Science or a related discipline as well
as experience in some of the following:
* » High performance software development in C/C++ or FORTRAN.
* » Program analysis and transformation
* » Compiler design and implementation
* » Computer architecture / digital circuit design.
* » Working with scientific software code in HPC application
domains such as computational finance, seismic processing,
life sciences or fluid dynamics.
We've looked at a many resumes (sadly, resume writing is frequently creative writing) and interviewed many candidates without finding many people who even come close to our standards.
To be fair, that is a VERY specific skill-set. I'm sure you know that, but I'd bet that most of the people with that skill-set are already employed somewhere (someplace like Intel would be my best guess) using it and not as amenable to moving on as, say, Johnny Rails-Developer would be.
It sounds like a very interesting and challenging gig, though. Out of curiosity, what's the salary for something like this?
Did you notice that the skills/experience bullet list said "some". We are interested in abilities, insight, and basic knowledge. As part of the candidate evaluation process we test candidates on their performance with a substantial design and implementation project, one that we believe should be fairly simple for someone with good programming skills and a good knowledge of algorithms and design. Most fail, and most of the rest perform badly. Even at the coder level, we see people who are unable to translate written descriptions into code. It is really depressing.
Maxeler is an interesting and challenging gig. Maxeler takes clients "known to work" code running on a cluster of machines (typically several thousand) and accelerates it by as much as a factor of 250X using a combination of hardware and software. Salaries depend upon skills and performance and are competitive with those elsewhere.
Comments
Conversely, why is it so hard to find good developers, people who really know something? For example, here's what we are looking for at Maxeler (http://www.maxeler.com):
We've looked at a many resumes (sadly, resume writing is frequently creative writing) and interviewed many candidates without finding many people who even come close to our standards.To be fair, that is a VERY specific skill-set. I'm sure you know that, but I'd bet that most of the people with that skill-set are already employed somewhere (someplace like Intel would be my best guess) using it and not as amenable to moving on as, say, Johnny Rails-Developer would be.
It sounds like a very interesting and challenging gig, though. Out of curiosity, what's the salary for something like this?
Did you notice that the skills/experience bullet list said "some". We are interested in abilities, insight, and basic knowledge. As part of the candidate evaluation process we test candidates on their performance with a substantial design and implementation project, one that we believe should be fairly simple for someone with good programming skills and a good knowledge of algorithms and design. Most fail, and most of the rest perform badly. Even at the coder level, we see people who are unable to translate written descriptions into code. It is really depressing.
Maxeler is an interesting and challenging gig. Maxeler takes clients "known to work" code running on a cluster of machines (typically several thousand) and accelerates it by as much as a factor of 250X using a combination of hardware and software. Salaries depend upon skills and performance and are competitive with those elsewhere.
Good question, I guess its open for debate