Why is it so difficult to find PCB design skills in today's graduates? We're having the greatest position in the world, lots of freedom, cool stuff to work on etc. and can't seem to find good talent to hire. There are plenty of EE applicants but very few have PCB design experience.
Is PCB design completely overlooked by today's schools? Most EE graduates have no clue and those that do have some clue have designed a 1 layer board with 5-10 components.
PS. btw. I hope this is allowed by forum rules, here's our position, in case anyone is interested: https://virtalica.com/company/#careers
Graduates of what? High school? Tech school? Two-year associates degree? Four-year baccalaureate degree? Some sort of PCB designer apprentice program? Maybe your company is looking in the wrong places. PCB layout is VERY specific to the type of circuitry and components the PCB supports. You learn this art
from experience DOING it, not usually from some professor. There is plenty of help available online for self-directed (i.e. product oriented) learning and PCB design programs, such as Altium, offer free video tutorials. And there is, of course, old fashioned printed texts with equations and explanations on how to layout impedance-controlled high-frequency signal paths.
Perhaps you should hire a "young turk" who is willing to learn the business. Even amateur hobbyists (I include myself in that group) have access to free PCB design software. And there are still a few PCB manufacturers remaining in North America. Perhaps you can ask some of them for hiring recommendations.
I am a "retired" electrical engineer (BEE, 1978, University of Dayton), but I spent many years working as a technician in the Electronics Laboratory at the University of Dayton Research Institute (UDRI), starting in 1967 after a four-year enlistment tour in the U.S. Air Force. When I started my "technician career" I always had the goal of eventually becoming an electrical engineer. UDRI provided that opportunity with full-time employment and 100% paid college tuition. I did have to purchase my own text books, but I didn't need the GI Bill to finance my education, nor did I have any student debt after graduating.
Along the way, I introduced the other technicians in the lab to printed circuit board design. Up until I was hired, they made "one off" circuits using point-to-point wiring on terminal posts, manually swaged into fiberglass boards. This works for most transistors and passive components, but fails when trying to wire up complex analog and digital circuits with dozens of closely-spaced terminals.
Back then (1970s), our PCB layouts were done with "puppets" placed on velum, with connections defined by crepe tape. The only practical boards were single-sided boards with through-hole components. We made these ourselves using Bishop Graphics for artwork and Kodak A-B lithographic film, contact-exposed and developed for photosensitized board exposure. Pre-sensitized boards were contact-exposed through developed negative A-B film under a strong photoflood lamp. Exposed boards were then "developed" in trichlorethylene, which dissolved the unexposed portions, and then etched in heated ferric chloride to separate the circuit traces from the copper-laminated fiberglass board. Lastly, the board through-holes were drilled with carbide drills and components mounted.
There were four or five of us technicians and we all became self-taught in PCB layout design. As circuits we worked with became increasingly more complicated, we eventually switched to commercial PCB houses but retained the manual artwork layups because UDRI claimed they could not afford to purchase any PCB design software seats. And, no, none of the classes I took for my engineering degree offered any instruction in how to layout PCBs. The Internet wasn't yet in existence but the campus library was well-stocked.
The "smart folks" were earning six-figure incomes, designing and laying out integrated circuits using razor swivel-knives to cut out sections of rubylith film, a very slow and error-prone process. The rest is history and a LOT of EDA (Electronic Design Automation) software. These folks were of course "electrical engineers," with a few solid-state physicists thrown in for seasoning, because you cannot afford to design integrated circuits by "trial and error" and expect to sell them at a profit. Similar to real estate, it's all about yield, yield, yield instead of location, location, location. At that time, California or Texas or Arizona) were THE places to be for bleeding-edge electronics development. However, we eventually gave that all away.
So, fast forward to the 1980s... I had finally graduated after ten long years of part-time study and went to work for a military contractor who happened to be dedicated to using Digital Equipment Corporation (DEC) minicomputers. Minicomputers were on the way out by then, but almost no one in management realized it. Personal computers (PCs) were on the rise, with lots of "IBM clones" available everywhere from Asian sources. Software applications (apps) for PCs were everywhere, some available as shareware and others outright stolen, even if "protected" by various means that were eventually "hacked" by some kid living in his mother's basement. Never underestimate a ten-year old with a limited budget but loads of time on their hands.
I tried to "push" the practicality and cost advantages of the early PCs and was told to (basically) shut up and use DEC minicomputers. Embedded PCs were a relatively new concept in the 1980s, but that is what I was initially hired to design... a man-portable data acquisition system that used an embedded Intel 8085 to record reflectance data on digital tape. This replaced a "kludge" that used a Radio Shack audio tape recorder and a 110 baud acoustic modem. Minicomputers, such as the DEC PDP-11 series, were mostly used for "scientific" applications of this type. Some were embedded and dedicated to the task, but some were not.
Throughout the 1980s until my job was terminated in 1990, it was an uphill fight with zero company support for any sort of "toy" computer. I was able to design and build or upgrade only a few systems with embedded IBM PCs (the real thing usually, not often a clone) before, eventually, the company that hired me lost their lucrative (and exclusive) classified "black" contracts with the intelligence and reconnaissance community. Sensor platforms moved away from the chemistry of film-based imagery and fully embraced hyperspectral electronic imagery.
In the end, a chemical engineer and I were the first to be "let go" as the company was bleeding cash to pay their expensive software analysts, most of who had TOP SECRET - Special Intelligence (TS-SI) security clearances. It didn't work. More lay-offs followed and eventually the company went out of business. Those with the high-level clearances just went to work for a similar company that is still successful today.
If you think it is expensive to hire a competent PCB designer, who can roll multi-layer boards with 60 GHz signal paths, try paying for the extended background checks required of that employee who needs a TS-SI clearance. The employer, not the Government, pays for this. And the clearances are immediately invalid if the employee quits to join the competition, unless the competition happens to also have government contracts that require those clearances.
Designing 9-12 layer PCBs and routing 20-40 Gbps is not trivial.
Not only is the design not trivial, this level of PCB design is virtually impossible to service in the field. So, new PCB designs must work the first time, every time. Re-work and re-design will kill off any profit or expected ROI. I have owned several desktop PCs with ASUS motherboards that eventually became obsolete. It was not even worth the effort to "salvage" the CPU and supporting chip sets because a brand new, state-of-the-art motherboard is virtually "free" compared to the silicon that populates it, or the support hardware that makes it a personal computer.
So, where do you go to find high-performance circuit board designers? Try India and the Peoples Republic of China. I hear they work cheap and today turn out excellent products, although they mostly don't care about intellectual property rights. Or hire a "hands on" technician who isn't afraid to learn new skills. Make sure you provide the tools this technician/PCB designer will need to succeed, and pay them a decent salary. That's decent, not competitive. You probably cannot afford to pay a competitive salary.