My company's doing apprenticeships; take somebody with no CS education, not even a college degree, and teach them to code.
It's an interesting model, I remember folks on HN calling for more companies to try it since it seems to have had some success in Germany, but I predict a lot more security bugs and unmaintainable code in the near future.
I'm not college educated and I've been coding since I was about nine in different capacities. I was inspired by two engineers my dad worked with that were the same; one of them was a SWE and the other was a Systems Engineer. I ended up dropping out before I hit any of my CS courses, so I basically did a repeat of high school.
The stuff you need to know for most jobs can be learned through books (DS&A); everyone, including grads, learn to actually code on the job. Systemic thinking and breaking problems down into manageable chunks is harder to train for; this is where I think something that's akin to apprenticeship could really help. At least the way I view it, and maybe I'm wrong, is that in the early 2000's, much less the 90's, there weren't many CS or CE schools - much less accredited ones that followed CAC standards. If your company is doing this then they're just getting back to the roots of what a computer programmer used to be.
We've done this, and it's worked out pretty well - all of our apprentices were dedicated to learning to write software, about half were offered jobs at the end of the apprenticeships, and half of those accepted.
Remember that they're not just mashing code straight to the main branch; they're apprentices, so other engineers paired with them, others read their pull requests, etc. It wasn't a free-for-all, nor should it be.
> It's an interesting model, I remember folks on HN calling for more companies to try it since it seems to have had some success in Germany, but I predict a lot more security bugs and unmaintainable code in the near future.
German here. The secret sauce behind the Duales System is that it's, as the name suggests, a split system - one part of the training is at government-run schools ("Berufsschule"), and the other part at the company that trains and pays you. And since the curricula are virtually the same across the schools, even if they're a bit outdated, they still produce decent graduates.
We don't make glamorous "unicorns" in the US style as we lack both the financials (aka enormous amounts of dumb money floating around waiting to be invested - remember, we don't have 401k, we have rolling-contribution pensions so no need for that) and the culture (we're very risk averse)... but our "Mittelstand" has produced hundreds of small and medium size companies that are world leaders in their respective market [1], often dominating their market with 70-90% [2]. And the foundation of that, especially for the older companies, isn't academics, it's trades, training and apprenticeship. Many of the Mittelstand companies, you enter in your youth and remain there for the rest of your life.
Our pride as a nation, our role models, is not a few people who struck it right to become multi-billionaires, our pride is the millions of people working for the Mittelstand and the consistently high quality of the stuff they produce. Boring, but wildly profitable and very, very resilient.
PS: You actually might know some of these things our tradespeople built. BMW/Audi/Mercedes/Volkswagen cars, MAN trucks, Rheinmetall, KMW and ThyssenKrupp military hardware from tanks to the massive Panzerhaubitze 2000, Diehl's IRIS-T anti-air defense, Heckler & Koch/Walther guns, anything with "Siemens" on it built before Siemens fell to MBA shenanigans, all developed and prior to globalization also built in Germany. And a lot of it, especially the military tech, is up to par with what the US military builds - for IRIS-T SLM and PzH 2000, the Ukraine war shows that they are even better to some experts.
One thing we sadly lost was pharmaceuticals - up until the 60s-70s, Germany used to be the "apothecary of the world" [3], but we lost that to India and China.
This only works if you hire enough really good engineers and have them spend most of their time mentoring, and most companies aren't willing to accept the senior engineers individual contributions on paper dropping to zero.
I disagree. Actively mentoring, like what this demands, is more supervisory. I do it with college interns all the time. It's not suddenly different because the pipeline doesn't originate with college. The process generally goes:
1. In weekly one on ones we may discuss a topic. I ask them to apply that topic.
2. They pick up sprint tasks and look to apply the knowledge they've gained.
3. They may ask some questions along the way; it's important that other engineers are also available for question asking - the same way peers may depend on each others knowledge.
4. You peer review the outcome in a PR.
Rinse and repeat.
I'll add I end up having to do this with everyone if they're fresh to industry or came from a place with poor standards for code writing and/or problem solving.
https://www.yearup.org/
A non-profit that I've volunteered with,that explicitly trains folks from alternate pathways to place them at entry-level positions, with support from certain corporate entities, ranging from training support to directly hiring via internships/apprenticeships.
In Germany, and Portugal (where I went through similar system), you still need to finish high school, and there is a programming related curriculum anyway.
The part-time work is like doing the labs part.
Also at the end of it, you can still go to the university if feeling like it. I did so.
Going through technical school was a secure way to have a job, in case the university exams weren't good enough for the engineering degree, which by the way is mostly state sponsored on this side of the Atlantic.
Amazon has a few different programs to retrain people for tech jobs, such as non-tech Amazonians, people separating from the military, under-represented people, etc.
It's an interesting model, I remember folks on HN calling for more companies to try it since it seems to have had some success in Germany, but I predict a lot more security bugs and unmaintainable code in the near future.