Job Skills Switching careers to software development

  • Thread starter Thread starter diegzumillo
  • Start date Start date
  • Tags Tags
    Careers Software
AI Thread Summary
Transitioning from academia to software development can be challenging, especially without prior experience in the field. Individuals with programming skills, particularly in machine learning, are encouraged to leverage their academic background and apply for tech jobs. Familiarity with key machine learning concepts and terminology is essential, as well as understanding the business context in which models are applied. Networking and showcasing projects on platforms like GitHub can enhance job prospects. Success stories highlight that it is possible to secure a developer position while integrating academic interests into a new career path.
diegzumillo
Messages
177
Reaction score
20
Hi all

In my country, and with the pandemic aggravating affairs, an academic career seems unlikely for me at the moment. It's what I have been preparing for, I finished my PhD and started looking into post doc positions nearby, but no luck so far. So people advised me to try becoming a software developer. I have programming skills, I have tinkered with machine learning, so it should be possible. But in practice things are a little harder. I have no prior experience working in the field, I know little of the terminology they use etc.

Anyone who made this transition, care to give me some advice?
 
Physics news on Phys.org
I didn't make the transition but am familiar with books on ML.

The best ones I've seen are:

- Hands-on Machine Learning with Scikit-Learn, Keras and Tensor Flow by Geron

https://www.amazon.com/dp/1492032646/?tag=pfamazon01-20

- 100 pg Machine Learning book by Burkov

https://www.amazon.com/dp/199957950X/?tag=pfamazon01-20

An honorable mention would be the Machine Learning Cookbook.

The most popular ML development language right now is Python although people do use Matlab, and Julia with code converted to Golang and Java for production level code.
 
I went physics -> actuarial -> data science. It has worked out very well for me.

I think Jedishrfu has a great list. Let me add to that a couple of others: 1) Elements of Statistical Learning, 2) Machine Learning by Murphy, and 3) any good undergraduate optimization book (e.g. Optimization by Chong and Zak). Remember - every statistical and machine learning model has an optimization process at its core!

I would also suggest grabbing some basic books on classical statistics and work through them. You'd be surprised how often analytics candidates with strong scikit learn/tensorflow/etc. experience get crushed by interview questions like "Describe what a confidence interval is" (the answer is trickier than it is given credit for). Both Google and Facebook will ask such questions for some roles.

Another piece of advice I'd give is that 100% of data science and machine learning is about people. It's about human beings making decisions. The models are being used for something, and without a strong understanding of the business or operational context, the work is wasted. A question I like to ask people is "Why would a model with a very low cross-validated error lead to spectacularly bad decisions?" Try to understand how these tools are actually used in business, why they can sometimes add spectacular value, and why they fail so often.
 
  • Like
Likes jedishrfu
diegzumillo said:
In my country ...
Is it a secret what that country is? It might matter. For example, if it's India in particular you would be up against a HUGE number of people already in the computer field. Probably the same with China. That doesn't mean it's necessarily a bad idea but it would certainly be something to be aware of.
 
Since you already seem to have at least some academic experience related to data science and software development, my opinion is that you don't necessarily need to do anything except start applying to the kinds of jobs you want in the tech field.
 
phinds said:
Is it a secret what that country is? It might matter. For example, if it's India in particular you would be up against a HUGE number of people already in the computer field. Probably the same with China. That doesn't mean it's necessarily a bad idea but it would certainly be something to be aware of.
In other posts, the OP has indicated that he/she/they are from Brazil. See the following (including your response).

https://www.physicsforums.com/threa...ents-outside-of-academia.994239/#post-6398713
 
StatGuy2000 said:
In other posts
OK, so NOW I know and I appreciate your wanting to be helpful but such information, when not presented in a thread is not helpful if you can't remember it (and I can't remember ANYTHING). He should either have it in his profile or have mentioned it as part of his question.
 
You are right. It did not occur to me this was important information.

But if anyone's curious, I did get a job as a developer! :) Good benefits and decent starting salary. I still feel bad about leaving the academia, and hopefully I can still integrate physics research in my life.

For anyone in a similar situation here is what I did: I uploaded every project I had to github and made them public. In my resume, since I didn't have any work experience, I just described some of the projects. The resume has a link to my linkedin and my github account.

I start only in October though.
 
  • Like
Likes Locrian, PeroK, phinds and 1 other person
diegzumillo said:
But if anyone's curious, I did get a job as a developer!
Glad to hear it. Congratulations.
 
  • Like
Likes diegzumillo
  • #10
phinds said:
Glad to hear it. Congratulations.
Ditto!
 

Similar threads

Replies
18
Views
4K
Replies
2
Views
2K
Replies
7
Views
2K
Replies
2
Views
4K
Replies
19
Views
3K
Replies
2
Views
3K
Back
Top