What changes between a senior engineer's job and an architect's, and what stays the same?
Show the full answer Hide the answer
What is being tested
Whether you understand progression as scope and judgement rather than as accumulated technology knowledge.
What changes
Scope. From one system to several, then to a domain, then to an organisation. The unit of concern grows and the level of detail necessarily falls.
The proportion of the job that is communication. Substantially higher. A good decision that is not adopted has no value, so persuasion, writing and facilitation become load-bearing rather than supporting skills. This is what strong engineers most often find uncomfortable.
Working through others. At senior levels, developing judgement in ten engineers has more effect than making excellent decisions personally. The measurable output becomes indirect.
Time horizon. From this quarter to the next two or three years, which means designing for changes that have not been requested and constraints that do not yet bind.
Comfort with uncertainty. More decisions with less information, and more of them irreversible.
What stays the same
Technical credibility, which the role entirely depends on. Architects who stop reading code lose it within a couple of years, and influence, mentoring and direction all rest on it.
Judgement about trade-offs, which does not become less technical.
Curiosity, and the habit of finding out rather than assuming.
What is actually rewarded
Outcomes — systems delivered, problems avoided, costs removed, teams made faster. Judgement, demonstrated by decisions that aged well. Making others effective. Owning the hard problems nobody else will take. Communication.
What is rewarded less than people expect
Knowing many technologies. Being right in arguments. Producing documents. Being present in every discussion.
What to do deliberately
Seek scope, not title — take responsibility for something larger than assigned; the title follows. Write, publicly or internally, which builds credibility and forces the thinking. Deepen the fundamentals, which do not expire. Get feedback on judgement, not just on output. Keep building. And record what you were wrong about, which is how judgement gets calibrated and which almost nobody does.