After mentoring more than 100 students and junior engineers, I’ve noticed that most graduation project problems start on day one — with the idea. Either it’s too big to finish, or too small to impress anyone. Here’s how to find the middle.
1. Start with a real problem, not a technology
“An app using AI and blockchain” isn’t a project. “Students at my university can’t easily find which labs are free” is. Start with a problem you or people around you actually have. It makes every later decision easier, and it gives you a much better story for your defense.
2. Write the core in one sentence
Your project should have one main thing it does well:
“A system that lets clinic receptionists book appointments and sends patients a WhatsApp reminder.”
If you can’t describe the core in one sentence, the idea is still too vague.
3. Split features into “must have” and “nice to have”
List every feature you want. Then be honest: which ones does the core sentence actually need? Build those first. Everything else goes on a second list you only touch once the core works end to end.
A small project that works completely is better than a big one that half works.
4. Pick tools you can learn in time
Your team’s current skills matter more than what’s popular. A well-built project in Python or JavaScript beats an unfinished one in a framework nobody on the team has used. Learning one new tool is fine. Learning five is how deadlines get missed.
5. Plan backwards from your deadline
Take your submission date and work backwards: testing and fixes, report writing, defense practice. Most teams underestimate these. Whatever time is left is your real build time, and it’s usually less than you think.
6. Prepare your defense from the start
The committee will ask why you chose your approach and how parts of the system work. Keep short notes on your decisions as you go. By the end, you’ll understand your own project deeply — which is the whole point.
A note on getting help
Getting help is normal, and good mentors make you better. But your project should be yours: you should be able to explain every part of it. That’s what I focus on when I work with students — teaching and guiding, never doing the work for you.