AI
AI didn't kill the junior role. It changed the job description.
The tasks that used to teach juniors their craft are now done in seconds by AI tools. That makes fundamentals more important, not less, for engineers and for go-to-market teams.
There is a popular argument that AI coding assistants and AI research tools will eliminate entry-level jobs. I understand why people make it. The tasks we used to give new engineers, such as boilerplate, simple bug fixes and first drafts of tests, are exactly the tasks AI does well. The same is true for junior marketers and business development associates, whose early work was often list-building, research and first-draft copy.
But I do not think the junior role is disappearing. I think its job description has been rewritten, and most training has not caught up.
What junior work used to teach
Boilerplate and simple fixes were never valuable in themselves. They were how people learned. Writing a hundred small functions taught you how a codebase fits together. Researching a hundred accounts taught you what a good-fit customer looks like. The work was an apprenticeship disguised as output.
When AI takes over that output, the apprenticeship has to be replaced deliberately, or it disappears.
The new junior job
A junior engineer's job now looks less like producing first drafts and more like judging them:
- Specifying what needs to be built clearly enough that a tool, or a colleague, can build it.
- Reading and reviewing code they did not write, and spotting what is subtly wrong.
- Testing and verifying behaviour, rather than assuming generated code does what it claims.
- Debugging the failures that tools cannot explain.
- Integrating pieces into a system that works as a whole.
For junior go-to-market people it is similar. Research and first drafts come fast; the value is in choosing the right accounts, knowing when the AI-written message is off, and having real conversations with customers.
You cannot review what you do not understand. That makes fundamentals the most important thing a junior can learn.
Why fundamentals matter more now
To judge AI output, you need a model of what good looks like. An engineer who understands memory, concurrency and data structures can tell when generated code will fall over under load. A marketer who understands positioning and statistics can tell when a tool's campaign analysis is nonsense. Without those foundations, AI makes people faster at producing things they cannot evaluate, which is dangerous.
What companies should do
Keep hiring juniors, and hire them for judgement, curiosity and communication rather than raw output. Pair them with both AI tools and senior people. Give them review responsibilities early, with a senior reviewing their reviews. And measure them on outcomes, not on lines of code or number of emails sent.
How the Fellowship reflects this
This thinking shaped both Fellowship tracks. The engineering track spends its first module on computer systems and algorithms, then teaches Rust, which forces a real understanding of ownership and concurrency, before four weeks of AI and ML engineering. We also teach AI-assisted development explicitly, including when not to use it. Some exercises are done without AI on purpose. The Go-To-Market track starts with strategy and statistics before automating anything, and requires human review in every AI workflow Fellows build.
The junior role is not going away. It is becoming more demanding, and more interesting. We want our Fellows to be ready for the new version of it.
