I'm a third-year BSIT student who enjoys building applications and websites, but I'm worried that I rely too heavily on AI when programming. I often use prompts to generate code, then read the explanations and try to understand what each part does. However, I'm concerned that I won't be able to compete for jobs after graduation if I can't write enough code independently.
I have experience with HTML, CSS, JavaScript, Android Studio using Java and Kotlin, PHP, XAMPP, and phpMyAdmin. I also have some experience using GitHub and collaborating with group members. Seeing other students' impressive résumés has made me feel behind, so I'd like advice on what skills or habits I should focus on improving.
I'd also appreciate advice about UI/UX tools. I know the basics of Figma, but I find it overwhelming and don't understand how to turn a Figma design into CSS for websites or layouts and XML for Android projects.
3 Answers
The résumés you see are usually polished highlights, not a complete picture of someone’s abilities. Focus on becoming comfortable with fundamentals: data structures, debugging, version control, HTTP, databases, and writing clean code. Make two or three projects that solve a real problem, document what you built, and be prepared to explain your design choices. AI-assisted development is likely to be common, but employers will still value people who can reason about problems, verify generated code, and work independently.
Feeling overwhelmed by Figma is normal. Start with simple wireframes and practice the basic ideas: spacing, alignment, typography, colors, and reusable components. Figma usually won’t convert a finished design perfectly into production CSS or Android layouts. Treat it as a visual specification, then manually reproduce the design using HTML/CSS or Android views and constraints. Compare your implementation with the design, adjust it, and repeat. That process will teach you both design and frontend layout skills.
AI can be useful, but it shouldn’t be the thing carrying your projects. Try building small features from scratch first, without asking AI to write the solution. If you get stuck, ask for an explanation, a hint, or help debugging rather than requesting the entire program. Then close the AI response and recreate the solution yourself. You should be able to explain every important line and modify the code without depending on the tool. Building a few complete projects and understanding them deeply will help more than collecting a long list of technologies.
That makes sense. I’ll try writing more of the code myself and use AI mainly for hints, explanations, and debugging instead of having it generate everything.