On Wednesday, 24 June 2015 at 13:04:29 UTC, Martin Drašar wrote:
Dne 24.6.2015 v 14:15 Chris via Digitalmars-d napsal(a):
Sure, this smells like a job description tailor made for
someone they already have in mind. However, I wonder what
"experience" means. "Yeah, I've played around with Java and
C++" or "I'm really into the two languages and know them very
well". People who do research for their PhD in computational
linguistics are often _familiar_ with programming languages
but not necessarily to a degree that allows them to write real
world production code, because it is very hard to excel in
programming, data algorithms, NLP and academic research
(literature review, writing the PhD) at the same time.
If you let researchers write production code, you are asking
for troubles. And that comes from someone who is doing
research, is a programming enthusiast and had programming as a
day job for some time.
That's what I mean. You either program or you do research. If you
do research, all programs are proofs-of-concept (with papers and
journals in mind).
If I made such job add, the most important for me would be the
ML experience and publication skills. This is something you
can't teach in a short time. As for the rest of it - it would
be good enough that the candidates would not hurt themselves
while using those tools and languages. At least there will be
less unlearning to do...
Martin
