Scope it to one semester, not one career
The most common rejection reason is over-ambition. A topic that needs a research lab, proprietary data or a year of training time will be sent back. Pick a problem you can demonstrate end-to-end on a laptop within four months.
Anchor it to a real, local problem
Panels respond well to projects with a visible user. Attendance systems, crop advisory tools, hospital queue management and campus logistics apps all read as practical work rather than a tutorial repeat.
Write your problem statement in two sentences a non-technical person can follow. If you cannot, the topic is still too vague.
Prepare the viva while you build
Keep a running log of every design decision and every alternative you rejected. Ninety per cent of viva questions are simply 'why did you do it this way?', and a log turns that into an easy answer.