After graduation, what actually works.
Straight answers about which paths are open, what each job is really like on a Tuesday afternoon, and what to learn in what order. Including the parts that are harder than they were, because pretending otherwise helps nobody.
No salary figures, no placement rates, and no recommended paid courses. Those are the three things this kind of advice invents most freely, they are the easiest to get wrong, and being wrong about them costs you money and years. Every path below says when it was last reviewed, so you can judge how current it is.
Paths
-
Cloud and DevOps
Keeping systems running and making deployment boring. Harder to enter straight from graduation, but unusually durable once you are in.
5 stages -
Cybersecurity
Defending systems that were not built to be defended. Much more process and analysis than the popular image of it suggests.
5 stages -
Data and Machine Learning
Turning messy data into decisions. Much more data cleaning and much less model building than the course adverts suggest.
5 stages -
Design and User Experience
Making things people can actually use. A portfolio decides everything here, which makes the degree matter less than anywhere else.
5 stages -
Product and Business Analysis
Deciding what should be built and why. The main route into tech for people who are analytical but do not want to write code all day.
5 stages -
Quality Assurance and Testing
Finding out what breaks before users do. Still one of the most realistic first jobs in tech, and a well-worn route into several others.
5 stages -
Software Engineering
Building and maintaining the systems other people depend on. Still the widest door into the industry, but the entry bar has moved.
5 stages
Three things that apply to all of them
Evidence beats credentials
Every path here is judged on what you can show. A finished, deployed project, a merged pull request, a usability study, a test suite that runs — each of these outweighs a stack of course certificates, and the gap has grown, not shrunk.
The bottom rung has moved up
Routine entry-level work — simple code from a clear spec, first-line triage, manual test scripts, production design work — is exactly what has become cheapest to automate. That is genuinely harder for new graduates. The work that remains is the work requiring judgement, which means the fundamentals matter more now, not less.
Sideways entry is normal
Several of these are rarely a first job. Support, testing and operations are ordinary routes into product, DevOps and engineering, taken by people who then did well. Entering adjacent and moving across is a strategy, not a failure.