Re: job postings on the web site
Jarrod Millman <millman <at> berkeley.edu>
2008-04-05 03:17:56 GMT
On Fri, Apr 4, 2008 at 7:27 PM, Robert Kern <robert.kern <at> gmail.com> wrote:
> I suspect that most of the time, the page will be mostly empty.
> Previously, I've stated that it would be permissible to post job
> announcements to the numpy and scipy mailing lists if they have a
> [JOB] marker in the subject line (and no one objected, so it's
> essentially official policy). For the level of activity that I'm
> estimating, it seems that that would be more appropriate. Fixtures
> like job pages don't really work unless if there is a de minimis
> amount of activity; if there aren't usually any jobs posted, no one
> looking for a job checks it. I suspect that you would get more quality
> hits by posting to the general Python Job Board than a numpy-specific
> one.
>
> http://www.python.org/community/jobs/
I agree with Robert on this; I don't think having a job posting page
would be of any use (I don't mind if you want to set it up; but there
are plenty of useful things you could do for the project instead). I
think that sending an email to the mailing lists and posting on the
Python Job Board make more sense.
I have hired several programmers to work on or with NumPy/SciPy over
the last several years. Finding good people is hard and personal
networking is the most effective solution. This isn't a NumPy/SciPy
problem; finding good technical people is hard in all fields (e.g.,
sysadmins, network admins, DB admins, scientific programmers,
informatics programmers). I tend to use services like Monster or
Craigslist, but they are rarely (if ever) useful. Personal contacts
always work best. For NumPy/SciPy programmers, I would recommend
trying the following (at least, these are the types of things I have
found the most useful):
1. Become an active member of NumPy/SciPy. Follow the mailing list
discussions and code development. Contribute code and documentation
yourself. Attend the SciPy conference and let people know what you
need help with.
2. Become active in your local Python group. This will put you in
touch with local Python programmers at least.
3. Form a local NumPy/SciPy community. This has been, by far, the
most useful thing I have done. Host coding sprints, meetings, or
hands-on workshops (e.g., invite Eric andTravis or Fernando and John
Hunter to give one of their introduction to scientific programming
with Python). Each of these activities bring in different groups of
people (the sprints and meetings will increase your exposure to the
NumPy/SciPy developers, while the workshops will introduce you to
developers/scientists who are starting to use Python for scientific
development). An important side-effect of these activities is that it
also gives you the opportunity to develop local NumPy/SciPy expertise
as well as help build the community. When I host sprints or meetings
at Berkeley or elsewhere, I often meet new people who I may work with
in the future.
I would be interested in hearing what other people have done to find
good NumPy/SciPy programmers.
By the way, Joe, did you post both jobs to the list? I only remember
one job posting of yours. Did I miss the other?
--
--
Jarrod Millman
Computational Infrastructure for Research Labs
10 Giannini Hall, UC Berkeley
phone: 510.643.4014
http://cirl.berkeley.edu/
Indeed PyMVPA itself might be
not the best basis to jump off for scikits.learn since we see
PyMVPA more of a framework, not a toolbox. As I said earlier, on the
other hand some ideas might be borrowed from PyMVPA. Please don't take
me wrong -- if you think that PyMVPA looks like a good starting point to
develop on -- that is cool, you are welcome to the team. On the other
hand, if you develop .learn independently we will be interested into
migrating some of functionality into wider-audience .learn and just rely
on importing .learn internally within PyMVPA
May be we could have some quick (or not so quick) chit chat (voice via
skype or any other voip + VNC screen) some time next week. There is
additional discussion happening on exppsy-pymvpa mailing list which you
are welcome to join as well
RSS Feed