Specialization in BE/ME that will be in demand for next 15 Years

It is not any specialization. All skills have a half life , ie how long will it take after 1/2 of the skill that you have is not useful. Generally the half life of many computing skills is about -3 years, Mechanical engineering is about 5–6 years.

The one item that will be in demand in 10 years is your

“ability to take a vague and ill defined problem, identify the underlying (as opposed to stated) need and actual (as opposed to stated) constraints (social thinking skill) , convert it into a proper hypothesis/need statement (critical thinking) and then craft a solution (using your combination of special skills and experience) and then defend your solution to your peers or others (communication and critical thinking and debating skills)”

If you dont have this, no specific skill will help you survive.

So rather than trying to anticipate what will be useful (we are very poor at judging this), you need the ability to learn a new skill and upgrade throughout your life: for this you need to have a strong conceptual foundation (ie ability to reuse ideas in many different contexts) and a very healthy curiosity, boldness and discipline to learn new things.

5 years ago I was convinced that MATLAB was going to dominate the engineering world…..well Ive now switched everything to Python, who knew….

I dumped 20 years of matlab expertise. But it was not a waste, I used my matlab skills to abstract the ideas and use them in Python. Ie from matlab I figured out “what I needed to do”, then, thanks to Stack Overflow I learnt how to do it in Python (thanks folks and I am perfectly willing to take the abuse if you teach me something). Eventually I started thinking “natively” in python and found better ways.

As computers and AI develops, specific skills will become less relevant.

I *hope* that eventually AI will be used to assist humans with the routine tasks (ie the star-trek model) and not the other way round (the Matrix/Terminator Model)

Hopefully AI would lead at worst to cyborgs and not robots.

-Arun Srinivasa,PhD UC Berkeley