70 FYP Ideas for Computer Science Students (2026)

Ezitech profile photo

Ezitech

Careers & Internships article by Ezitech — 70 FYP Ideas for Computer Science Students (2026)

Your final year project is the one thing every interviewer will ask about. It also decides what you build for the next two semesters. This page lists 70 FYP ideas for computer science students, grouped by field, with a tech stack and difficulty level for each.

Every idea is sized for a team of 2–3 students over two semesters. Difficulty is marked Easy, Medium or Hard, so you can match an idea to your team’s skills. The last list covers ideas that solve Pakistani problems. These are easier to defend in a proposal because the problem is local and real.

Jump to: AI & Machine Learning · Generative AI & Agents · Computer Vision · Web Development · Mobile Apps · IoT · Cybersecurity & Blockchain · Data Science · Pakistan-Focused Ideas

Want a shortcut? Use our FYP idea generator to filter these 70 ideas by interest, team skills and difficulty, or get a random pick.

Ideas for your degree: software engineering (BSSE) · IT (BSIT) · AI · cyber security. Need a name for your project? See 60 FYP title examples. Looking for money to build it? See FYP funding in Pakistan.

How to choose the right FYP idea

Pick an idea you can finish, demo and explain, not the one that sounds most impressive. Check it against these five questions before you write the proposal:

  1. Is the problem real? Name a person or business that has this problem. If you can’t, the proposal defense will be hard.
  2. Can you get the data? A common reason AI projects stall is that the dataset doesn’t exist or isn’t labelled. Find it on Kaggle or Hugging Face, or plan how you will collect it, before you commit.
  3. Does it fit two semesters? Cut the idea down to one core feature that works end to end. Add extra modules after the mid evaluation.
  4. Does it match your team? Split the work by skill: backend, frontend or mobile, and the AI model or hardware.
  5. Will it help you get hired? Build in the stack you want a job in. A deployed React and Node.js project helps a web developer more than a desktop app nobody can open.

AI and machine learning FYP ideas

An AI project scores best when the model solves one clear problem and runs inside a working app, not just a notebook.

# Idea What it does Tech stack Difficulty
1 Student dropout early-warning system Predicts which students may fail a course from attendance, quiz marks and LMS activity, with a dashboard for teachers Python, scikit-learn, Flask, React Medium
2 Fake product review detector Flags suspicious reviews on e-commerce listings from review text and reviewer behaviour Python, BERT, FastAPI Medium
3 Personalised learning path recommender Suggests courses and a study order based on a student’s skills and career goal Python, collaborative filtering, Django Medium
4 Credit risk scoring for microfinance Predicts loan default risk and explains each score so a loan officer can trust it Python, XGBoost, SHAP, Streamlit Hard
5 AI resume screening and ranking Ranks CVs against a job description and shows why each one matched Python, spaCy, sentence embeddings, React Medium
6 Electricity load forecasting Forecasts hourly demand for a building or feeder to plan load management Python, LSTM or Prophet, Grafana Hard
7 Road accident hotspot predictor Maps high-risk road segments from accident and weather data Python, GeoPandas, Leaflet Medium
8 Speech emotion recognition for call centres Detects caller emotion from voice clips to flag calls for quality review Python, librosa, CNN Hard

Related: How to Become an AI Engineer · AI Internship at Ezitech

Generative AI and AI agent FYP ideas

Generative AI projects are the easiest to demo in 2026, but only if the output is grounded in real data and checked for mistakes.

# Idea What it does Tech stack Difficulty
9 University handbook chatbot (RAG) Answers student questions from the prospectus and rules, with the source page cited Python, LlamaIndex or LangChain, Chroma, React Medium
10 Bookkeeping agent for small shops Reads invoices and receipts, sorts expenses and drafts a monthly summary Python, LLM API, OCR, PostgreSQL Hard
11 Code review assistant for student repos Reviews GitHub pull requests and explains bugs and style issues in simple words Node.js, GitHub API, LLM API Medium
12 Voice interview practice bot Asks role-specific technical questions, listens to answers and gives scored feedback Python, Whisper, LLM API, React Medium
13 Contract and agreement simplifier Turns rental agreements and contracts into plain summaries and flags risky clauses Python, LLM API, PDF parsing Medium
14 Multi-agent research assistant One agent searches, one summarises and one fact-checks, producing a cited report Python, LangGraph or CrewAI Hard
15 Quiz generator for teachers Builds MCQs from lecture slides or PDFs at three difficulty levels and exports them to Google Forms Python, LLM API, Google Forms API Easy
16 MCP server for campus systems Lets AI assistants read timetable, attendance and library data through the Model Context Protocol TypeScript, MCP SDK Medium

To keep API costs near zero, run an open-weight model locally with Ollama during development. Related: Open-weight vs API LLMs · MCP Explained · Ezitech AI Bootcamp

Computer vision FYP ideas

Vision projects give the strongest live demo, since the panel can watch the camera work in the room.

# Idea What it does Tech stack Difficulty
17 Face recognition attendance Marks class attendance from one classroom photo and blocks photo-of-a-photo spoofing Python, OpenCV, InsightFace, Flask Medium
18 Desi food calorie estimator Identifies dishes like biryani or daal from a photo and estimates calories per serving Python, YOLOv8, Flutter Hard
19 Number plate recognition for parking Reads Pakistani number plates at the gate and logs entry and exit times Python, YOLO, EasyOCR Medium
20 Sign language to text Translates Pakistan Sign Language gestures into Urdu or English text in real time Python, MediaPipe, LSTM Hard
21 Helmet violation detector Spots motorcycle riders without helmets in CCTV footage Python, YOLOv8 Medium
22 Retail shelf stock monitor Detects empty shelf space from store camera images and alerts staff Python, YOLO, FastAPI Medium
23 Handwritten answer sheet grader Reads handwritten short answers and suggests marks against an answer key Python, TrOCR, LLM API Hard

Start with a public dataset, then add a few hundred images you collect yourself so the model works on local roads, food and lighting.

Web development FYP ideas

A web project stands out when it is deployed on a live URL with real users, not only run on localhost during the defense.

# Idea What it does Tech stack Difficulty
24 Hostel booking platform for small cities Students find, compare and book verified hostels near campus, with reviews and a map Next.js, Node.js, MongoDB, Maps API Medium
25 Local services marketplace Book plumbers, electricians and tutors, with ratings and in-app payment Laravel, MySQL, Vue.js Medium
26 Online exam system with proctoring Timed exams with shuffled questions, tab-switch alerts and webcam checks MERN stack, WebRTC Hard
27 Multi-vendor store with courier tracking Sellers list products, admin takes a commission, orders sync with a courier API Next.js or Laravel, PostgreSQL Hard
28 Clinic appointment and e-prescription portal Online booking, doctor schedules, digital prescriptions and SMS reminders React, Node.js, PostgreSQL Medium
29 Live collaborative whiteboard Several students draw, type and chat on one shared board in real time React, Socket.IO, Canvas API Medium
30 Alumni job referral portal Alumni post openings, students request referrals, the admin verifies both sides Django, PostgreSQL Easy
31 Event ticketing with QR check-in Sells tickets for campus events and checks guests in by scanning a QR code Next.js, Node.js, QR library Easy

Related: MERN Stack Developer Roadmap · Web Development Internship at Ezitech

Mobile app FYP ideas

A mobile FYP carries more weight when the app is on the Play Store or shared as a test build that real people have installed.

# Idea What it does Tech stack Difficulty
32 Blood donor finder Matches patients with nearby donors by blood group and location, with emergency alerts Flutter, Firebase, Google Maps Easy
33 Offline-first field survey app Collects survey forms with no internet and syncs them when a connection returns Flutter, SQLite, REST API Medium
34 Women’s safety SOS app One tap shares live location with trusted contacts and starts an audio recording Flutter or Kotlin, Firebase Medium
35 Medicine reminder for elderly users Gives dose reminders in Urdu voice and alerts a caregiver when a dose is missed React Native, Firebase, text-to-speech Easy
36 Campus lost-and-found with photo matching Matches found items to lost reports by comparing photos Flutter, Firebase, image embeddings Medium
37 Student carpool app Matches riders on the same route to campus and splits fuel costs Flutter, Node.js, Maps API Medium
38 Hostel expense splitter Tracks shared bills between roommates and shows who owes whom React Native, Firebase Easy
39 AR campus navigation Shows the way to classrooms and offices through the phone camera Unity with AR Foundation, ARCore Hard

Related: Flutter Developer Roadmap · Offline-First Apps for Pakistan · App Development Internship at Ezitech

IoT and embedded FYP ideas

IoT projects combine hardware, a backend and an app, so they suit teams with one electronics-minded member.

# Idea What it does Tech stack Difficulty
40 Solar system monitoring dashboard Tracks panel output, battery level and load, and alerts when output drops from dust or a fault ESP32, voltage and current sensors, MQTT, Grafana Medium
41 Smart irrigation controller Waters crops based on soil moisture and the weather forecast ESP32, soil moisture sensor, weather API Medium
42 Gas leak detector with auto shut-off Detects an LPG or gas leak, closes the valve and sends an SMS Arduino, MQ-5 sensor, GSM module, solenoid valve Easy
43 Appliance-level energy meter Measures each appliance’s usage and shows an estimated bill in an app ESP32, PZEM-004T, Flutter Medium
44 Cold chain monitor for medicines Logs fridge and transport temperature and alerts when it goes out of range ESP32, DS18B20, GSM or LoRa Medium
45 Smart parking slot finder Sensors detect free slots and an app shows what’s available ESP32, ultrasonic sensors, Firebase, Flutter Medium
46 Water tank and pump automation Starts and stops the pump by tank level and logs daily water use ESP32, ultrasonic sensor, relay module Easy

Simulate the circuit in Wokwi or Tinkercad before you buy parts, and put the hardware budget in your proposal.

Cybersecurity and blockchain FYP ideas

Security and blockchain projects are judged on whether the threat or trust problem is real, so open your proposal with a real incident or use case.

# Idea What it does Tech stack Difficulty
47 Phishing link and scam SMS detector Classifies links and SMS messages, such as fake prize offers, as safe or phishing Python, scikit-learn, Chrome extension Medium
48 Network intrusion detection system Detects attacks in network traffic with a model trained on a public dataset like CICIDS Python, Scapy, scikit-learn Hard
49 Zero-knowledge password manager Encrypts passwords on the device so the server never sees them in plain text React, Web Crypto API, Node.js Hard
50 Web vulnerability scanner Checks a website for common OWASP issues like XSS and missing security headers, then writes a report Python, OWASP ZAP API Medium
51 Deepfake image detector Flags faces in images or videos that have been digitally altered Python, PyTorch Hard
52 Degree verification on blockchain Universities issue degrees on chain, and employers verify them by scanning a QR code Solidity, Ethereum testnet, React Medium
53 Medicine supply chain tracker Follows each batch from manufacturer to pharmacy to catch counterfeits Solidity or Hyperledger Fabric, Node.js Hard
54 Transparent donation tracker for NGOs Shows donors where each rupee was spent, recorded on chain Solidity, Polygon testnet, Next.js Medium

Only scan or attack systems you own or have written permission to test, and say so in your proposal’s ethics section.

Data science and analytics FYP ideas

A data science FYP is only as good as its data, so pick an idea where you can collect or download the data in the first month.

# Idea What it does Tech stack Difficulty
55 Pakistan job market analyzer Collects job postings and shows in-demand skills and salary ranges by city Python, Scrapy, Pandas, Streamlit Medium
56 Brand sentiment tracker Tracks public opinion about a brand across social posts in English and Roman Urdu Python, Hugging Face transformers, Streamlit Medium
57 Retail sales forecasting dashboard Forecasts weekly sales per product and suggests how much stock to reorder Python, Prophet, Power BI Medium
58 Smog and air quality predictor Predicts next-day AQI for a city from weather and pollution data Python, XGBoost, Streamlit Medium
59 University result analytics Shows which courses have high failure rates and how grades change by semester Python, Pandas, Power BI Easy
60 Customer churn prediction for ISPs Predicts which customers may cancel and the main reasons behind it Python, scikit-learn, SHAP, Streamlit Medium

When you collect data from websites, follow their terms of use and robots.txt, and anonymise any personal data.

FYP ideas that solve Pakistani problems

These ideas are hard to copy from a foreign tutorial, which makes them stronger in a proposal defense and more memorable in a job interview.

# Idea What it does Tech stack Difficulty
61 Roman Urdu toxicity classifier Sorts Roman Urdu comments into positive, negative and abusive to help moderate Pakistani pages Python, multilingual BERT, FastAPI Medium
62 Urdu Nastaliq OCR Converts scanned printed Urdu into editable text Python, Tesseract Urdu model or fine-tuned TrOCR Hard
63 Crop disease detector for farmers A farmer photographs a wheat, cotton or rice leaf and gets the disease and treatment in Urdu Python, TensorFlow Lite, Flutter Medium
64 Solar size and payback planner Suggests panel, inverter and battery size from monthly bills and shows the payback period React, Node.js Easy
65 POS for small shops with FBR invoicing Billing app that works offline and prepares invoices in FBR’s digital invoicing format Flutter or Laravel, SQLite Medium
66 Campus canteen ordering with Raast QR Students order and pay through a simulated Raast QR flow; the canteen sees live sales Flutter, Node.js Medium
67 Mandi price tracker Shows daily wholesale crop and vegetable prices by market and sends price alerts Python, Firebase, Flutter Easy
68 Urdu voice guide for public services Answers spoken Urdu questions like how to renew a CNIC, using official government pages Python, Whisper, LLM API Hard
69 Flood alert and relief matching app Shows river-level alerts from official sources and matches relief donors to needs by area Flutter, Firebase, Maps API Medium
70 Local wagon and bus route finder Crowd-sourced routes for local wagons and buses in cities that have no transit data on Google Maps React Native, Node.js, Maps API Medium

Related: Building AI Products for Urdu and Roman Urdu · FBR Digital Invoicing Integration · Raast Payment Integration · Agritech Software in Pakistan

What makes an FYP stand out to employers

Employers don’t grade your report. They open your link, watch your demo and ask what you built yourself. These five things make that conversation easy:

  • It’s live. A working URL, a Play Store link or a demo video anyone can open.
  • The GitHub repo is clean. A README with screenshots, setup steps and a short note on the architecture.
  • It has numbers. Model accuracy on a test set, page load time or the number of real users.
  • Someone actually used it. A shop, clinic or university department that tried it, with one line of feedback.
  • There’s a 2-minute demo on LinkedIn. A recruiter can check a short video in the time it takes to open a 100-page report.

More on this: Build a Developer Portfolio That Gets Replies

The typical FYP timeline

Most Pakistani universities run the FYP across the 7th and 8th semesters in these stages. Exact deadlines differ, so check your department’s FYP handbook.

  1. Choose an idea and a supervisor at the start of the 7th semester.
  2. Submit the proposal and present it at the proposal defense.
  3. Write the SRS, the software requirements document.
  4. Write the design document, covering architecture, database and UI mockups.
  5. Show a working core feature at the mid evaluation.
  6. Build and test the full system during the 8th semester.
  7. Submit the final report and demo it at the final defense.

Once you have picked an idea, the next three steps are documents, not code: the proposal format and what each section must contain, then the SRS, diagrams and final report, and finally the viva questions you will be asked.

Common FYP mistakes to avoid

Most FYPs that go wrong fail on scope and teamwork, not on technology.

  • An idea with no new angle. A plain library or hostel management system is fine only if you add something new, like AI search or a real user.
  • Too much scope. “Uber, Foodpanda and Daraz in one app” never ships. One core feature that works beats five half-built ones.
  • No data plan for an AI idea. If you can’t name your dataset in the proposal, change the idea.
  • Leaving documentation to the last month. Write the SRS and design document as you build, not after.
  • One person writes all the code. At the defense, every member must explain every module.
  • Buying a ready-made project. Panels often ask you to change the code live, and a bought project teaches you nothing you can use in an interview.
  • No Git, no backups. Push to GitHub from week one so a dead laptop doesn’t end your project.

Frequently asked questions

What is a good FYP idea for BSCS students?

A good BSCS FYP solves a real problem, uses data you can actually get, and fits into two semesters. An AI model running inside a working app, like ideas 1 to 23 above, is a strong choice in 2026.

Can I do my FYP in Flutter or the MERN stack?

Yes. Most departments accept any stack if the project is complex enough and your supervisor approves it. Flutter suits mobile ideas, and MERN or Next.js suits web platforms.

How many students are in an FYP group?

Usually two or three. Some universities allow one to four, so check your department’s FYP rules.

Can I build an AI FYP without a GPU?

Yes. Google Colab and Kaggle give free GPU time, and starting from a pre-trained model cuts training time a lot. Many generative AI ideas call an API or a small local model and need no GPU at all.

How do I write an FYP proposal?

A proposal usually covers the problem statement, objectives, scope, related work, methodology, tools, timeline and expected results. Keep the scope to what you can deliver, and name your dataset or hardware early.

Which FYP is best for getting a job?

The one built in the stack you want to work in, and live for anyone to try. Build a web platform for a web developer role, or a trained model inside an app for an AI role.

Build the skills your FYP needs

Most FYP teams get stuck on the build, not the idea. An internship at Ezitech lets you work on hands-on projects in web, mobile, AI and design before your defense, so you walk in having already shipped something.

Explore Ezitech Internships · AI Internship · Web Development Internship

We update this list every year before proposal season, so bookmark it and share it with your FYP group.

Leave a Reply