I'm in my third year of a BSIT degree and really enjoy building systems, especially apps and websites. Lately, though, I've been worried that I rely too much on AI when programming. I often describe what I want and let it generate code, then read through the explanation and try to understand each line. I'm learning, but I'm concerned about whether I'll be able to compete for jobs after graduation, especially when I see portfolios and resumes from students who seem much more advanced.
I'm familiar with HTML, CSS, JavaScript, Android Studio with Java and Kotlin, and have some experience with PHP, XAMPP, and phpMyAdmin. I also understand the basics of GitHub and have used it to share work with teammates.
What would be the best way to improve my skills and become more independent? Is using AI acceptable if I use it mainly for explanations, debugging, and guidance, or should I be writing more code from scratch? Also, are there beginner-friendly UI/UX tools or learning resources that can help me design app screens and then translate those designs into CSS or Android XML? I know about Figma, but I find it overwhelming and don't yet understand how to turn a design into a working interface.
4 Answers
Your current technologies are a reasonable foundation, but focus less on collecting tools and more on completing a few substantial projects. Build something with authentication, a database, validation, error handling, documentation, and deployment. Use Git with meaningful commits and write a clear README. Those projects will teach you more than copying many small tutorials, and they give you something concrete to discuss in interviews. Comparing yourself with polished resumes is also misleading because you rarely see the unfinished projects and gaps behind them.
You should take the concern seriously, but it doesn’t mean you are doomed. AI-assisted development is likely to remain common, while fundamental knowledge will still matter when generated code is incorrect, insecure, inefficient, or difficult to maintain. Practice regularly without assistance, review AI output critically, and make sure you can solve basic problems under time pressure. A useful routine is to attempt a feature alone, compare your solution with an AI suggestion afterward, and then explain the differences in your own words.
For UI work, start with simple wireframes rather than trying to master every feature in a design application. Sketch the screens on paper or use a basic template, decide what each screen should do, and then implement one screen at a time. In Figma, you mainly need frames, text, shapes, components, spacing, colors, and basic prototyping at first. CSS and Android layouts are not automatically converted perfectly from a design, so treat the design as a visual specification: measure the spacing, identify the hierarchy, choose consistent colors and typography, and recreate those choices with layout containers and reusable styles.
AI can be useful, but you should make sure it isn’t replacing the part where you develop problem-solving skills. Try building smaller projects without AI first, even if they take longer. When you get stuck, ask for an explanation or a hint instead of immediately requesting the complete solution. You should eventually be able to explain the code, modify it, debug it, and recreate the important parts yourself. Employers generally care more about whether you can build and maintain something than whether you typed every character manually.
That makes sense. I’m going to spend more time writing projects myself and use AI mainly when I need an explanation or help finding a bug.