Perhaps this has changed, given that we're in a semiconductor boom cycle now, but I have my doubts. When SMIC started trying to poach TSMC talent, the response was not to make pay competitive (or even half decent), let alone in line with the massive geopolitically-relevant value being created. No, the response was a mask-off legislative crackdown to keep the nerds in line. In the US, there were only two big employers, and they were definitely paying what they could get away with.
Anyone who is considering this career -- any career, but this one especially -- before you jump, please get the perspective of someone in industry who isn't trying to sell you a career.
Everything you're seeing in the comments is true.
Compensation is not that bad. Clearly, it'll not pay SW salaries - no engineering does. But if you're a fab person, you'll work long hours, be on call often (and you will get woken up often), and eventually will own a tool that you'll be responsible for, even when not on call.
Lots of abusive and pathological behavior, as well. And they often block internal transfers so you're basically trapped.
People with other skills (e.g. SW) get out. The rest are stuck, because they have, for example, a chemistry PhD and no other company will pay more.
See this thread from a while ago:
For some reason, I assumed that in order to be a chip designer you had to be as good at programming as a developer. Then I became friends with two intel chip designers/engineers and was surprised that neither knew much of anything about programming, despite both having multiple PhDs in their fields. That's when I began realizing that engineering and programming are 2 entirely different fields that have little overlap.
Folks who work on Chromium think little of those who wrangle wordpress themes
Things you can build and see and have immediate feedback.
That just doesn’t happen with chip design. Not that it can’t, and I’m sure there are examples (Woz’s paper and pencil circuit design come to mind) but it’s far from common, compared to coding.
I remember in 5th grade being asked to help another student write a choose-your-own-adventure style game in basic on an apple //e back in the “one per classroom if you are lucky” days.
I figured I was going to be a programmer as a profession but college came around and it was a coin toss for me between computer engineering and computer science. At the time, engineers made more money so I went with that. Doubled as an EE/CE. I was ok at the typical EE stuff, better at the little bit of programming we were exposed to (in pascal). I did very well in semiconductor physics, but I didn’t really get sucked into it. Got to a digital design class, and that hit the spot for me. I didn’t like plugging wires in on breadboards because it was entirely too frustrating trying to figure out which wire you got wrong, plus I’m colorblind and the shades of red and green insulation on wires are perfectly impossible for me to tell apart except under very very bright light. But I loved designing the synchronous digital logic.
It’s just not something you are super likely to randomly pick up on your own at an early age. And to be honest, even if I had been exposed to digital design back in grade school, I doubt it would have resonated. I could easily understand things like “GOTO 10.” Understanding clocked logic, or binary arithmetic probably wasn’t within my reach back then.
Nearly all of the EEs in my class had never programmed anything before the 2 classes in our curriculum. All the people that knew how to program went into CS (except for a few like me). So the few logic designers that come out of an EE program don’t seem very likely to be good at programming. Just different interests.
Both require deep domain knowledge that doesn't always overlap - but being able to do both makes you more valuable
Trust me, I've tried to talk them out of it and never succeeded.
I worked for a supplier for AMAT/Intel/Samsung etc. and was basically on-call 24/7. I had to carry two laptops with me at all times (my personal one and a work one).
I didn't care because I was young, and well compensated, but they decided that they'd cut my pay by a third by converting me to a salaried employee. The unpaid overtime (previously paid as a contractor) lost its luster quickly. Once I realized I was subsidizing bad management because I was filling the gaps left by often intentional shortcuts taken-- for free, I bailed. That took... two months I think.
Every person I interacted with in the semiconductor space was either an academic who was stuck in a lab 12+ hours a day or some line manager for Intel/IBM/TSMC/etc. that was on-call 24/7 if something went wrong. Both of those sounded terrible to me regardless of the pay.
So yeah I do software now.
Ended up as a power systems engineer, which is the complete opposite in terms of scale from semiconductors!
Terrible pay and terrible hours and long school and hard work is just irredeemable.
That was until I did an internship following a grad student doing work on solar cells. Every day was sitting 8+ hours in one room doing CVD then XRD. It was soul crushing. I realized the program was more of a PhD-prep program and most people I know who graduated and didn’t go for a PhD either 1) went into software or 2) work as some menial process engineer working the “assembly line” with no meaningful upwards mobility.
I know a surprising amount of people who came out of CNSE (without following up with grad school) and are now landscapers, waiters, or similar. The skillset is so specialized as to be almost worthless unless you can find the right role imo. Also the program there CNSE/SUNY Poly is absolute hot garbage, I advise active looking at this field to go into conventional EE, ChemE, or Materials.
In any commercial organization, there are tremendous pressures in high-investment, super high-risk projects and anyone can trivially find things that are wrong with the current system. The much more difficult challenge is showing an alternate path that is superior - based on the results it achieves. The times when you see the ugly behavior of people are times of desperation. Online commenters expend a lot of effort point out how to improve things by "proving" it by linking to random studies - but the more persuasive argument is by implementing those changes in the real world. "Talk is cheap, show me the code." ;)
This code is working great!
A mentor of mine hit his breaking point when they split/bankrupt/acquired his team to discharge pension obligations. Real nasty stuff. He said everyone was retiring, "good luck with the next node" (10nm) -- which I discounted as sour grapes on account of Intel appearing invincible at the time, but wow have the years cast that story in a different light.
Was that a joke? What's the point of doing two PhDs in the same field? That's like 12 years wasted. 1/8th of your life.
(Worked at Intel and along with fab/process folks for a bunch of years).
https://www.cgw.com/Press-Center/News/2014/Studios-Accused-o...
https://www.vanityfair.com/news/business/2014/04/apple-googl...
How about instead of idolizing people like this we charge them with felonies and put them in prison?
Pay was not so great as judged by the cars. The other thing was that the workforce was split - you were either old or young.
If you want to make money in semiconductors, find a place they are building fabs and become a plumber or electrician. Those are the folks who make money, and often there’s security clearance requirements that adds job security.
Is a foundry the only employer?
Last time I had close contact with the industry, digital design indeed was a bit more competitive and correspondingly it was the only one considered to have good career prospects -- but the offers were still between shabby and embarrassing next to entry level SWE, even outside the bay area.
This could have changed.
Then you have to manufacture. There is a significant cost for custom chips. Though once you get going the per chip cost is pretty low, thus the business lends itself to a few large players.
Points they told:
1. A mandatory PhD just to get at internship
2. Few years of unpaid internship needed to get a coffee porter job, before you are let anywhere close to the process
3. A postdoc is required to get a real R&D cred
4. 7-10 years in R&D sweatshop before any real promotion
5. Get very lucky
6. In 20-30 years of career, you can get to a VP level, where you can hold the company by the balls with critical knowledge. Then... and only then, you will look up to a 7 digit salary.
You wind up captive to a small number of employers in places that have few other employers. You have to be physically present--no online only. You will have weird hours because you have to slot into the fab plant openings. Your pay will be a small fraction of even a mediocre software developer.
I can go on and on.
There isn't a shortage of semiconductor personnel. We all fled.
Companies could pay people enough to come back. Like so many other fields, companies would rather do anything other than raise salaries.
Stay far away.
There's a normative and substantial laboratory aspect to ABET-accredited[1] undergraduate engineering programs.
From Criterion 7:
>> Modern tools, equipment, computing resources, and laboratories appropriate to the program must be available, accessible, and systematically maintained and upgraded to enable students to attain the student outcomes and to support program needs.
[1] https://www.abet.org/accreditation/accreditation-criteria/cr...
Most of the decent ones are expensive, and many still consider in/out of state tuition. There are a few good deals I've seen, highly dependent on being in-state.
The only "useful" ones you'll find tend to be CS, a handful of EE degrees exist online (no clue how that works) along with some other fields. Occasionally you'll find some in the natural sciences, but they tend to be BA's and explicitly for pre-law, pre-med, etc. I've seen maybe one online math degree and tons of social science ones, as well as "we made this major up to target to poor working people who will give us money".
At the end of the day, they exist, but are uncommon and probably not worth the price.
I researched going back to school online for years, but if I do it, I will probably do it in person (at least for the important years).
Forcing students to live on-premises also makes universities money through student housing and amenities. Universities have concerns that opening their programs to online instruction would lead to an exodus from the campus. You could have 70% of a 20,000 strong student body attend online, and all the expensive real estate of the campus would mostly be a waste of money.
- Even in my first attempt at college, I was never too concerned with the name/reputation of the school. My first go around I picked a state school that gave me a full ride, rather than going for the more prestigious options. So once I checked that the accreditation of WGU was indeed legit, that was good enough for me.
- Their admissions are based on your professional history, since they're set up to cater to working professionals, not high school graduates. (Heck, their mascot is the "night owl.") That means it didn't matter to them that my first attempt at college left me with a... bit of a "stained" transcript. That stopped me from getting into other online options, but my professional history spoke for me at WGU.
- I'm a huge fan of their competency-based education model. It isn't for everyone since the classes are self-paced and asynchronous, but it matches my personal abilities to self teach and to "binge" content. Plus, as soon as you enroll in a course you can take a pre-assessment that will unlock the final if you do well. For some of the courses I've spent over a decade in the workforce learning the content, so I was able to do the pre-assessment and then the final in a single day.
My current experience at WGU is short, but good. My term started June 1. I rocketed through a few of the classes that had test-based and code-based finals in the first week. Now I'm realizing just how out of practice I am with writing, because I'm in a pair of classes that have writing-based finals and I'm struggling more.
I don't personally know people that took their degree from WGU to another university for a Masters, but WGU maintains a list [0] of universities which have accepted WGU undergraduates into their graduate and/or doctoral programs.
[0] https://www.wgu.edu/alumni/career-support/education-and-prof...
The academia is a community and a lifestyle. The people who choose that lifestyle generally don't want to spend too much time teaching outside the community. Teaching is a lot of work, a lot of bureaucracy, and a lot of hard deadlines, and it's not particularly rewarding if the students are just names on screen. Maybe if people paid higher taxes and higher tuition fees, universities could hire teachers to "just work here" and pay them competitive salaries.
Majority of ppl here are after electrical eng.
Cs seems to be easy in compare
In CS, you may have spent thousands or tens of thousands of hours across many years learning and growing, and still being limited in your understanding of the field as a whole. It's daunting to explore other related fields just to find out that they each have similar levels of complexity, filled with professionals who've sunk similar levels of effort, time, and years of their life into their work (who are often equally struck by the complexity of your field.)
I find embedded really rewarding. A 20 year old embedded-C code base generally follows the same design patterns and coding conventions and styles that you'd use today, and don't feel "old" or like they need to be rewritten. A 7 year old JS code base, on the other hand, is largely outdated, and may be written in a nearly extinct framework by the time it is your turn to maintain it. If you get tired of giving up time on the weekends so you can learn a new framework, consider hopping over!
I'm exaggerating a little, but more power to those of you who suffer through it!
Ferrous Systems is making some good progress to make certification of Rust happen, through ferrocene[0] and other initiatives, but it will take time.
The only company I know using Rust in a sort-of embedded way is SoloKeys[1], using Rust to write the firmware of a hardware authentication token (similar to a yubikey).
Typically the real time space is very conservative on what types of changes are implemented, and that's largely because a lot of rt software exists on safety critical systems that need to be supported for decades.
Its pretty great to be honest, I think I would have a hard time going back to C/++
Several automotive companies have been hiring for Rust, though what exactly they’re using it for isn’t as well known publicly yet.
There have been a few aerospace projects. And at least one train project, IIRC?
In that kind of situation, you usually own the whole device. You don't have to worry about being removed from memory because the user wanted to run some hotel's booking app or whatever - the user doesn't have the option of doing that.
You may have some additional hooks that give you some control of whatever custom hardware that comes with the device. You may or may not have to drop down into native code to access those hooks.
You can't develop against a standard phone or tablet. You have to have your hardware to develop against.
So, yeah. It's the same... but it's different.
Other times it has more to do with the particular brand of, ahem, fertilizer, they use.
Sometimes it really is greener and your side is shit and it's time to make a change.
Is what?!
it's barely even engineering
Typically softeng roles require a degree in CS or equivalent, and many years of experience on top of that.
Then again, in some other companies "senior engineer" is somebody who did a 3-months bootcamp and 3 years of work. It's very unfortunate.
Ppl who I worked with went into more web like jobs after semico
But later I'd want to work as compiler engineer
Also, the mistake I made when I went to grad school: Semiconductors seemed to be a "new" field compared to the rest of EE. One of my undergrad professors said "They still haven't figured out what a standard textbook should contain."
In reality, from a research standpoint, it's a very mature field. Don't expect low hanging fruit. If you're going to focus on theory, expect it'll take a number of years of dedicated study before you get to the frontier. You'll need to know quantum mechanics and statistical mechanics, and some electromagnetics, just to begin studying semiconductor theory. Then a whole bunch of specialized solid state courses. Then you start studying the specific subtopics (reading key journal papers).
> But later I'd want to work as compiler engineer
Why are you wasting time with semiconductors...?
At least for the analog domain, Gray et al[1] has been the standard entry-level gateway drug for longer than I care to remember.
Semiconductor theory is more about the transistor and everything below it: Starting from quantum mechanics and statistical mechanics, derive the equations of electron/hole transport in a doped semiconductor, how the material's band structure impacts current flow, etc. The derivation of the transistor equations is often the end point.
Because this company does both, so after I get bored with current role and prepare better, then i'll try to switch teams
I didnt and still do not see easier way for me to get into comps (i may be naive), especially that i dont know cpp yet, which seems like must have in compilers world unless you go to e.g msft
Link to online degree: https://engineering.purdue.edu/semiconductors/degrees#online...