Six years of programming: the foundations that help me work with AI

After studying the artistic side of games, I trained in Unity and C#. On the course, we worked on a mobile game in which a character dived underwater and collected rubbish. It meant turning an easy-to-explain idea into specific behaviours within a program.

I also coordinated work with design and marketing. I discovered that finishing a project involved more than writing code: people needed to understand each other, share priorities and prepare deliverables. That experience still matters when an assistant speeds up implementation but the team needs to know what it is building.

My first junior developer job involved React. Alongside building interfaces and connecting services, I spent months fixing bugs: buttons that behaved incorrectly, inconsistent styling and functions that broke a screen.

Reading other people's code helped me recognise patterns and trace a problem. That practice offers a concrete way to learn today: when AI proposes a change, review what it modifies, which data it receives and the expected behaviour afterwards. Reproducing a bug gives you a reference for checking the fix.

In the recording, I describe periods at a large consultancy and a very small company. At one, I had to adapt to React Native with less support than expected; at the other, I worked directly with a backend colleague. These are personal experiences, not universal judgements about those kinds of company.

Both periods made me understand communication between the interface, services and data more clearly. Even if you do not implement every layer, you need to know what you are asking each one to do and how to investigate an unexpected response.

I began using AI assistants after working with code, documentation, colleagues and personal projects. That is why I distinguish learning to produce a screen from understanding the application behind it. A tool can suggest changes without automatically answering your questions about the system.

In the video, I mention frontend, backend, APIs and databases as part of that basic vocabulary. For practice, take a small function and draw its path: what the user enters, where it is processed and what returns to the screen. Then explain what you would check if that path were interrupted.

The journey includes changes of direction, jobs that ended and projects that did not go as expected. Summarising it in a video can make it look much more continuous than it actually felt. That summary is not a useful timetable for measuring someone who is starting out.

Your next step can be smaller and more useful: finish an interface, understand a request or document a resolved bug. AI can support that learning, but sustaining it means practising, checking and accepting that some concepts need more than one attempt.