FYP ideas that get you hired (and 3 that don't)

Your Final Year Project can do one of two things: get you hired, or become a line on your transcript nobody reads. The difference isn’t how fancy it sounds. It’s whether you can demo it in an interview.

An FYP that gets you hired has three things:

1. A real user or a real dataset. Not a toy example, something a real person would actually use, or real data you worked with. This is what makes the conversation interesting in an interview.
2. A deployed, live link. It runs. You can send a recruiter a URL and they can click it. “It works on my laptop” doesn’t count; shipping it is half the signal.
3. A real problem it solves. Even a small one. “I built X because Y was annoying for Z people” is a story. A feature pile with no problem behind it isn’t.

Three to avoid, because everyone builds them and recruiters have seen them a thousand times:

• ❌ An attendance / management system — the default FYP, says nothing about you.
• ❌ A simple chatbot wrapper — wrapping an API in a UI isn’t a project anymore.
• ❌ An e-commerce clone — a tutorial you followed, not a problem you solved.

Actual FYP ideas worth building. Pick one, keep the scope small, and make it real (a real user/dataset, deployed, on GitHub). These are directions, make them local and specific to a problem you’ve actually seen:

AI / ML (highest demand right now):

• A RAG assistant over a real document set — your university’s prospectus/handbook, past papers, or a company’s policies: ask a question, get an answer with the source. Teaches embeddings, vector search, and LLMs, the exact stack AI jobs want.
• A resume ↔ job-description matcher that scores a CV against a posting and lists what’s missing (an ATS-style tool).
• A fine-tuned or prompt-engineered classifier on a dataset you collected, e.g. detecting fake/scam job posts, or classifying Urdu/Roman-Urdu sentiment.
• An AI agent that does one real task end to end — e.g. scrapes job boards and emails you a filtered shortlist, or auto-drafts replies.

Full-stack / web (very hireable if deployed):

• A tool that solves a real problem at your university — a past-paper archive with search, a society/event portal, a hostel or mess-management app students actually use.
• A small SaaS for a local business — appointment booking for a clinic/salon, inventory + sales for a small shop, an invoicing tool for freelancers.
• A marketplace or directory for a specific niche (local tutors, used textbooks, blood donors) with real listings.

Data / analytics:

• A dashboard on a real dataset — Pakistani weather, cricket, electricity load-shedding, or data you scraped yourself, with genuine insights, not just charts.
• A price-comparison or tracker that pulls real data (e.g. laptop/phone prices across local stores) and alerts on drops.

Automation / IoT (if that’s your lane):

• An automation that saves someone real hours (a reporting pipeline, a WhatsApp/email bot for a small business).
• A focused IoT project with a working demo (smart attendance with a real sensor, a home-energy monitor) — hardware demos stand out if they actually run.

The pattern across all of these: a real problem, real data or a real user, and a link that works. Keep it small, ship it, and put the code on GitHub with a clear README.

One honest rule: one small FYP that’s live and that you can explain end-to-end beats ten impressive-sounding ideas you never shipped. Build the small thing, deploy it, and walk into interviews with a link.