Many thanks to Rachael Brown, Tushar Nagananda, John Nicholls and Tim Rawling for contributing insights to early drafts of this article.
The things that make an engineer great have been largely unchanged throughout history:
- First principles. An intense understanding of structural behaviour, including the ability to do a back-of-the-envelope calculation and land within 10-20% of the thorough answer.
- Intuition and judgement. The ability to sketch a solution and know what looks right and what doesn’t. It’s built through experience - pattern recognition accumulated across hundreds of different problems.
- Communication. Translating complex technical realities for non-technical stakeholders.
The pathway to developing all three is under serious threat. Numbers one and two in particular - and three depends on having one and two.
Until recently, the way an engineer built experience was through grinding out designs.
I remember in my early years building spreadsheets from scratch to calculate Super-T hogs, writing pages of baseplate design checks, doing simple strut-and-tie designs by hand. That work was tedious. It was inefficient. But it forced you to engage with the underlying behaviour. You built intuition as a side effect of just doing the job.
The old path was accidentally rich with lessons. The new path has accidentally stripped them out.
When calculations are performed by someone - or something - other than the engineer, how do you develop a feel for what works? When your role shifts to coordinating disciplines and managing clients, when getting into the weeds isn’t cost-effective, how do you build genuine understanding?
There’s an implicit deal at the heart of how our profession has always worked: the firm provides the work, the work provides the learning.
Engineering firms win big projects. Those projects require detailed calculations for many, many elements. But it’s not cost-efficient to have your $350 senior engineer - the one leading the project and shaping its direction - crunching out detailed calculations.
So we hire someone fresh out of university for half the price. The firm can now increase their profit margin on projects, or race to the bottom in bids, their choice.
And the graduate gets valuable experience. They’re practising engineering. Making mistakes, which is fine given their chargeout rate. Developing and becoming an asset to the engineering community.
Junior engineers develop almost as a byproduct of projects getting delivered. Firms get cheap labour. Juniors get on-the-job training. Eventually, everyone benefits from a well-trained engineer.
Win, win, win.
So, what’s threatening this deal?
I spoke to someone at a large multinational consultancy recently who mentioned they offshore much of their engineering work to India. At first I was surprised. Now I’m surprised more places don’t do it.
We’ve already largely offshored drafting. Engineering consultancies are fundamentally about labour arbitrage. So the logic is sound: keep a local engineer to interface with the client, manage other disciplines, and run the whole thing. Have a team in a lower-cost country do the grunt work of putting the design on a page and checking it holds together.
It’s the same play as hiring a graduate engineer, just taken to a more logical extreme.
Will we really be surprised when AI takes even those offshore tasks?
The difference is that these engineers we offshore the work to, and AI eventually, aren’t developing to run projects locally, for the most part. We get the short-term cost benefit but miss the long-term development benefits. We already complain there aren’t enough great senior engineers on the market, though there are a host of reasons for this.
So the implicit deal is breaking down. No longer do grads get development while firms get a reduced rate.
And this is not just because of offshoring and AI.
Remote and hybrid work has quietly made things much worse in this specific respect - and I say that as someone who values the flexibility it provides. Mentorship has always been largely informal and ambient. It happens over someone’s shoulder, in the car on the way to a site visit, in the conversation after a design review. That kind of incidental transmission of knowledge is nearly impossible to replicate on a Teams call.
WFH didn’t just make mentorship harder to schedule - it made the most valuable kind almost invisible.
This isn’t an argument against flexible work. It’s an argument that we haven’t yet figured out how to replace what we lost when we moved away from the office full-time.
So we now have multiple simultaneous pressures all hitting the same point: offshoring removes the technical grunt work, AI accelerates that process, and hybrid work erodes the informal mentorship that used to fill the gaps. No single one of these trends is necessarily fatal. Together, they’re a serious problem.
It’s not that young engineers are lazy or less capable. The industry has slowly dismantled the scaffolding they’ve climbed to competence and not put up a new one.
So where does the development burden go?
A few possibilities.
It shifts to the individual
Engineers self-direct their own learning outside billable hours through deliberate practice, self-study, and using AI as a Socratic tutor - something that asks “what do you expect the answer to be?” before running the calculation.
Some engineers will thrive under this model. Many won’t. And the inequality between the self-motivated and technically shallow will increase.
The best engineers have always been the ones who are self-motivated. This will continue to be true, however I believe it is an unsatisfactory industry response.
It shifts to universities
Firms want someone who joins at a practitioner level rather than a graduate. Universities respond by offering a master’s degree that does what grad programs used to do - except now instead of being paid to learn, the engineer pays to learn.
Engineers still need practical work-based experience that a university degree can never replicate.
Engineers used to begin as apprentices to master engineers. Formal education requirements came later - first through technical institutes and trade schools, then through university degrees. The bar has only ever moved in one direction.
Firms treat it as a competitive advantage
Some firms decide that producing great engineers is a competitive advantage and build real internal training programs, perhaps apprenticeships and mentorships.
They accept a short-term loss on paying young engineers to learn and develop but receive a long-term gain by creating a culture and environment that encourages the individual to stay with the firm and eventually deliver value far above their development cost.
I suspect most firms already do this calculus and decide the long-term gain doesn’t justify the short-term cost - especially when a well-trained engineer might leave before they can pay it back, though there are ways to mitigate this.
Third-party platforms fill the gap
Experienced people in the field recognise the gap and develop platforms and training resources where young engineers can learn and develop, for a fee, though considerably cheaper than a university.
This includes structured learning tools, simulated design environments, AI-driven tutors.
There are already people starting in this space with mixed success. Expect this market to get highly fragmented and specialised.
Right now, nobody is clearly stepping up to own this. This is an institutional problem. It requires an institutional response.
Companies will act in their self-interest and optimise for shorter-term profit. That’s not a criticism - it’s just how businesses work. But someone needs to be thinking about the profession on a longer time horizon. I’d love to see Engineers Australia and other engineering bodies put serious thought into this: what does engineering development look like in a world where the old system is becoming less relevant? What are the standards, structures, and incentives we need to build?
Because the consequences of poor training don’t show up immediately. They surface years later in failed projects, bad decisions, and assumptions nobody thought to challenge. By the time we notice, it’s too late to fix the cohort we failed.
What happens if in 20 years we have no great engineers?
Or - what if we already don’t, relative to a generation ago?