The best engineer you know thinks "platform-first"
Some thoughts on the characteristic of the best among us. TL;DR at the bottom.
The Situation
You've just joined the XYZ Finance team, your first task is to handle payments from the investment maintenance fees, customers always pay them using their credit cards. You sit down, design the solution, develop, deploy, et voila. The company can now charge customers their investment maintenance fee via credit card.
Three months go by and another fee comes up, "fund consulting fee". A smart engineer looks at your code and understands they can deliver really fast if they just create an enum or something and fork the process. The company can now charge customers their "fund consulting fee" via credit card.
Another two months go by and someone from the commercial department closes a deal with another company to manage the portfolio of every employee on their payroll. We must now charge every other fee, as well as a "B2B Portfolio Management Fee", and the company wants to pay all of this using cryptocurrency, or sometimes SWIFT deposits. Engineers sit down and increase the forks once again, but now not only for fees, but also for payment methods.
A couple of rounds of this "logical time save" later, and there are hundreds of incidents popping up, no one knows who owns what, and no one wants to fix anything. This is moment executives dread to get to: "We need a thousand man-hours to rewrite our fee charging engine. This will affect our roadmap."
The visionaries
I've been lucky enough to have met at least a dozen brilliant software engineers during my career, and it's not surprising that they tend to have their own set of specialties: some are more tech-savvy than others, some are more keen to understanding the business, others thrive in organizing and improving people, maybe some are just plain generalists and can solve mostly anything that you throw at them.
There is a single characteristic though, that I've spotted in every single one of them: they always approach a solution in a platform-first manner.
The best engineer you know thinks platform-first.
When presented with a problem, the solution is already popping up inside their head, that's not the issue at all. The question they're asking themselves is "how can I design this in a fashion that makes this scalable, governable, and/or distributable?", and the answer is most likely a platform.
These professionals can usually detect patterns across the organization that allow them to design things in actual reusable ways, after all, who hasn't seen two components sharing the same logic? The so called "platforms" are usually composed of multiple services, they are structural software that enables products to be built quicker, and to make sure the inputs and outputs (in reality the entire process as well) always follow a standard: valid, auditable, maintainable, and governable.
The vision is not the only requirement though. Platforms usually take more time to develop than the easy path, that's why I give our theoretical engineering super-hero another common quality: they are political creatures, and natural salespeople.
The best engineer you know thinks platform-first, and is also political and a natural salespeople.
It is not easy to sell the idea of a platform: "I have a tight roadmap, why should I spend hours in building something this big now?". The best engineers I've known could sell this idea to anyone, they know how to work their way around these conversations. They use realistic metrics, usually backed by other engineers who share the pain of having built the same functionality across many services before, and a long list of incidents that keep showing up along the way.
Many might have the same mindset as these people, but sometimes lack the history. They usually are composed of founders and early joiners, the company "old timers", deserving to be listened to because of their past glory. The accumulated knowledge of having developed a plethora of the companies products makes them easier to trust, and after the conversation is over, the only argument remaining is: "where are we getting all the people to work on this?"
I think this argument was the hardest one to overcome a couple of years ago, there simply weren't enough qualified people around to build these complex systems. AI helps us with this: with a well defined specification, software that would take weeks or months to be developed can be put to test in a couple of days.
"Well, if AI is around, what is the difference between this theoretical genius and any other software engineer?". That's easy, they already know what they want, and they really know how to ask.
The best engineer you know thinks platform-first, and besides being political and a natural salespeople, they are the most efficient AI users.
I'm glad I get to learn from these people.
TL;DR
The best engineers I've met try to build a platform before the product, they are also very political and can sell their ideas through realistic returns, and (because we can't ignore it) use AI efficiently to help thinks get off the ground.