Hot Topics and Skills in DSP industry

Dear all,

Hi, I am new to this forum.  I am a graduate student in signal processing
and still thinking whether to join academia or industry upon graduation. 
I have a few questions concerning the current DSP market, and I'd like to
gather insights from the experienced DSPers here.

1. What is the hottest topics in DSP industry right now?  Distributed
computing, sensor network, and ultrawideband are very popular in academia,
and I wonder if they are also of interest in the industry?

2. What are some useful skills outside school that I should learn to be
competitive in the market?  In school they only teach Matlab and some
simple DSP assembly lang... I feel that they are not enough.

3. This may be quite controversal: It is a better idea to start the career
at small companies as a senior employee, rather than a trainer at large
companies?  What is a good strategy for a fresh graduate (without working
exp) to start the career in the indsutry?

Thanks so much!
C J Lam


		
This message was sent using the Comp.DSP web interface on
www.DSPRelated.com

"clam" <jethrolam@gmail.com> wrote in message
news:WOidnTU4m7PswArfRVn-vg@giganews.com...
> Dear all,
>
> Hi, I am new to this forum.  I am a graduate student in signal processing
> and still thinking whether to join academia or industry upon graduation.
> I have a few questions concerning the current DSP market, and I'd like to
> gather insights from the experienced DSPers here.
>
> 1. What is the hottest topics in DSP industry right now?  Distributed
> computing, sensor network, and ultrawideband are very popular in academia,
> and I wonder if they are also of interest in the industry?

None of those are hot topics in my world, but then I work in audio.  I don't see
many questions on those topics in this forum either.  Maybe others have a better
perspective on this...

> 2. What are some useful skills outside school that I should learn to be
> competitive in the market?  In school they only teach Matlab and some
> simple DSP assembly lang... I feel that they are not enough.

Basic "nuts and bolts" programming skills are very useful (C, C++, and
assembly).  In most of my DSP projects, the algorithms aren't the only thing you
need to program.  There is also a lot of control code, setting up DMAs,
buffering, serial ports, etc.  Try implementing an actual algorithm that deals
with real-world I/O on a DSP evaluation board and you will get some experience
with the kind of topics I'm talking about.

Digital hardware design and troubleshoot skills are also very useful as well.
In the real world, the hardware won't always work perfectly, and you need to
figure out if problems are hardware, software, or both and then solve them.

> 3. This may be quite controversal: It is a better idea to start the career
> at small companies as a senior employee, rather than a trainer at large
> companies?  What is a good strategy for a fresh graduate (without working
> exp) to start the career in the indsutry?

(I assume you meant "_trainee_ at large companies"?)  I think a company who
hired a fresh graduate with no experience as a "senior employee" would be making
a big mistake!  Graduates need some time and usually assistance to come up to
speed and be productive employees.  I started my career at a small company, but
in a junior position.  There were several other more experience engineers who I
could go to for help and questions.  That worked well for me, and may for you as
well.  An internship is also a good way to get started.

clam wrote:
> Dear all,
> 
> Hi, I am new to this forum.  I am a graduate student in signal processing
> and still thinking whether to join academia or industry upon graduation. 
> I have a few questions concerning the current DSP market, and I'd like to
> gather insights from the experienced DSPers here.
> 
> 1. What is the hottest topics in DSP industry right now?  Distributed
> computing, sensor network, and ultrawideband are very popular in academia,
> and I wonder if they are also of interest in the industry?

The hottest topics that I run across in industry is actually being able 
to make things work.  Expect that unless you're doing real bleeding edge 
work you will be using techniques and theory that is anywhere from 10 to 
100 years old.  The only places where the real avant-guard stuff gets 
done is startups and large corporations (and the former are trying to 
get bought by the latter).

My success has always come from my ability to turn theory into practice, 
not my ability to generate new theory.  This means not only being able 
to make a prototype work in the lab, but being able to make a system 
that works in the lab, in the field, in the hands of uneducated 
customers, and through years-long production runs.
> 
> 2. What are some useful skills outside school that I should learn to be
> competitive in the market?  In school they only teach Matlab and some
> simple DSP assembly lang... I feel that they are not enough.

No, it isn't.  Figure that a knowledge of C is going to be an 
entry-level requirement, and many places will either require or look 
favorably upon C++.  Military contractors may well want to see ADA 
instead, and may view C/C++ experience with suspicion, so be ready to 
explain that you're willing to learn.

Unless you're going to be doing _only_ algorithm development you will be 
writing and maintaining software.  To do this well you need to know how 
to write good production-quality code, and how to maintain it.  Even if 
you are only writing algorithms you should learn how version control 
systems work, and learn enough about software quality assurance that you 
can play a positive role in the process.
> 
> 3. This may be quite controversal: It is a better idea to start the career
> at small companies as a senior employee, rather than a trainer at large
> companies?  What is a good strategy for a fresh graduate (without working
> exp) to start the career in the indsutry?

I did well by starting at smaller companies.  I think that if you have 
your sights set on a larger company you should try very hard to get in 
as an intern before you graduate.  This may be too late for you now, but 
if you have another year to go then start trying to get your foot in the 
door.
> 
> Thanks so much!
> C J Lam
> 
> 
> 		
> This message was sent using the Comp.DSP web interface on
> www.DSPRelated.com


-- 
-------------------------------------------
Tim Wescott
Wescott Design Services
http://www.wescottdesign.com

>Hi, I am new to this forum.  I am a graduate student in signal
processing
>and still thinking whether to join academia or industry upon graduation.

>I have a few questions concerning the current DSP market, and I'd like
to
>gather insights from the experienced DSPers here.
>
>1. What is the hottest topics in DSP industry right now?  Distributed
>computing, sensor network, and ultrawideband are very popular in
academia,
>and I wonder if they are also of interest in the industry?

Distributed Computing and DSP IMHO has got nothing to do with each other.
But to answer your question about hottest topic, I would say it is
consumer devices or end customer(or close end customer) boxes in
communication, personalization, entertainment like audio, video, music and
speech.In other words-Cell Phones.

>2. What are some useful skills outside school that I should learn to be
>competitive in the market?  

Three Important skills:

1] Debugging 
2] Debugging 
3] Debugging 

>3. This may be quite controversal: It is a better idea to start the
career
>at small companies as a senior employee, rather than a trainer at large
>companies?  What is a good strategy for a fresh graduate (without
working
>exp) to start the career in the indsutry?

I have never seen it through this prism at all and to be honest I dont
like this line of thought very much. To me the goal is to trasport
yourself from wherever you are to wherever you want to go. Provided you
know how to recognize both places. It doesnt matter whether you are at a
small company or a large one, what matters is the fire in the belly.

--Bhooshan
		
This message was sent using the Comp.DSP web interface on
www.DSPRelated.com

bhooshaniyer wrote:

> Distributed Computing and DSP IMHO has got nothing to do with each
> other.

They have quite a bit to do with each other if you're talking high-end
multi-dimensional processing --- some things just can't be done on a
single (multi-core) computer.

> Three Important skills:
>
> 1] Debugging
> 2] Debugging
> 3] Debugging

I'd rephrase this as:

1) Problem-solving
2) Problem-solving
3) Problem-solving

> To me the goal is to trasport
> yourself from wherever you are to wherever you want to go. Provided
> you know how to recognize both places.

That can be a HUGE ask.

> It doesnt matter whether you are at a
> small company or a large one, what matters is the fire in the belly.

I gotta agree with that.

Ciao,

Peter K.

bhooshaniyer wrote:
-snip-

> 
> Three Important skills:
> 
> 1] Debugging 
Finding and fixing bugs at the specification level is 1/10th as costly 
as fixing them there if they're found in the code.
> 2] Debugging 
Finding and fixing bugs in a code review is 1/10th as costly as fixing 
them during system integration.
> 3] Debugging 
Finding and fixing bugs during system integration is 1/10th as costly as 
fixing them during manufacturing rollout -- which is 1/10th as costly as 
fixing them after the product has been released to customers.

Finding bugs should be done like voting in Chicago -- early and often.
> 
-snip again-

> 
> I have never seen it through this prism at all and to be honest I dont
> like this line of thought very much. To me the goal is to trasport
> yourself from wherever you are to wherever you want to go. Provided you
> know how to recognize both places. It doesnt matter whether you are at a
> small company or a large one, what matters is the fire in the belly.

Yea verily.
> 
> --Bhooshan
> 		
> This message was sent using the Comp.DSP web interface on
> www.DSPRelated.com


-- 
-------------------------------------------
Tim Wescott
Wescott Design Services
http://www.wescottdesign.com

Thanks for all your input.  As some of you have said, I probably should
start writing control code, driver.. so that I can have something to
impress the employers. :)  
		
This message was sent using the Comp.DSP web interface on
www.DSPRelated.com

"clam" <jethrolam@gmail.com> wrote in
news:ULGdnVLmtKh-EgXfRVn-iA@giganews.com: 

> Thanks for all your input.  As some of you have said, I probably
> should start writing control code, driver.. so that I can have
> something to impress the employers. :)  

Take a look at the contests from ccink (circuitcellar.com). You would 
really impress your employers (or at least me :-) if you can present 
(one or several) projects you have done. It does not matter that the 
topics there are not DSP related. It counts that you can carry out 
whole projects yourself.

M.
-- 
Bitte auf mwnews2@pentax.boerde.de antworten.

Matthias Weingart wrote:

> "clam" <jethrolam@gmail.com> wrote in
> news:ULGdnVLmtKh-EgXfRVn-iA@giganews.com: 
> 
> 
>>Thanks for all your input.  As some of you have said, I probably
>>should start writing control code, driver.. so that I can have
>>something to impress the employers. :)  
> 
> 
> Take a look at the contests from ccink (circuitcellar.com). You would 
> really impress your employers (or at least me :-) if you can present 
> (one or several) projects you have done. It does not matter that the 
> topics there are not DSP related. It counts that you can carry out 
> whole projects yourself.
> 
> M.

That's a good first runner-up, right behind a summer job where you 
actually did things well enough to get paid.

-- 
-------------------------------------------
Tim Wescott
Wescott Design Services
http://www.wescottdesign.com

Tim Wescott wrote:

> That's a good first runner-up, right behind a summer job where you
> actually did things well enough to get paid.

Ya, no matter how much or how little money means to you personally,
future employers will ALWAYS judge you on how much you got paid, by
whom.

Ciao,

Peter K.