Advanced final year project ideas
Free advanced final year project ideas on Sefism for CS and software engineering students, each scoped for a single semester with a suggested stack and deliverables.
- 300 project ideas
Ideas and guidance, not finished projects
These are project ideas and scoping guidance, published free for students to use as a starting point. I do not build, write, or sell final-year projects, and I do not complete coursework for anyone. Take an idea, make it yours, and build it.
79 ideas · page 1 of 4
Multi-Agent Literature Review Assistant with Verified Citations for Students
Students writing a literature review often get chatbot answers full of papers that do not exist. Build a multi-agent assistant that plans a search, queries open scholarly databases, reads abstracts and open-access full texts, drafts a structured review, and refuses to cite anything it cannot resolve to a real DOI or arXiv record with a supporting passage. Then measure whether multiple agents actually beat a single agent on quality and cost.
Brand-Consistent AI Video Ad Generator with Automatic Checks and Disclosure Labels
Small brands want short video ads for Instagram, TikTok and YouTube but cannot afford a studio, and AI video tools produce clips that drift off-brand and are not clearly labelled as AI-made. Build a generator that turns a product brief and brand kit into short ads in several formats, then automatically checks every frame for logo, colours, text spelling and banned claims, and adds a visible AI label plus Content Credentials before export.
Human-Approved AI Agent for Filing and Tracking Citizen Service Requests
Filing a complaint or service request with a city or government office means finding the right portal, filling long forms, attaching documents and then chasing it for weeks. Build an agent that turns a citizen's plain description (in Urdu, Roman Urdu or English) into a correctly filled request, pauses for the citizen to approve every submission, then tracks status and reminds them to follow up. Built and tested against a realistic mock portal you create.
Drone Plant Stand Counting and Gap Mapping for Smallholder Maize and Sugarcane Fields
A few weeks after sowing, farmers need to know how many plants came up and where the gaps are, so they can resow in time. Build a pipeline that takes low-altitude RGB drone photos of a small field, stitches or tiles them, counts plants with a detection or density model, and produces a gap map and an estimate of plant population per acre. Compare it against hand counts in marked sample plots.
River Flood Forecasting with LSTM Models from Upstream Gauges and Satellite Rainfall
Train and evaluate machine learning models that forecast river discharge one to five days ahead for a few South Asian or African river basins, using upstream gauge readings, satellite rainfall and public hydrology datasets. Compare against simple persistence and published global forecasts, and report the lead time at which a warning threshold would have been crossed in past floods. This is the forecasting model, not a sensor network or an alert app.
Protocol-Locked Danger Sign Triage Assistant for Community Health Workers in Urdu, Bangla and Arabic
Build a phone assistant that walks a community health worker through a sick child visit in Urdu, Bangla or Arabic, where the triage decision comes from an encoded clinical protocol, never from a language model. The model only does what it is good at: understanding a spoken or typed answer in a local language and turning it into a protocol field. The project tests whether that split keeps the decisions safe while making the tool easier to use than a long paper form.
Acoustic Early Detection of Red Palm Weevil Larvae in Date Palm Trunks
The red palm weevil kills date palms from inside, and by the time a tree shows symptoms it is often too late. Build a low-cost clip-on sensor that records vibrations in the trunk, an embedded classifier that separates larval feeding sounds from wind, traffic and birds, and a dashboard for farm managers. Test it on recordings from infested and healthy trees with an entomology partner.
Citation-Checked Legal Question Answering over Pakistan Code and India Code
Law students and junior lawyers in Pakistan and India increasingly ask AI tools legal questions, and those tools sometimes cite sections that do not exist or misquote them. Build a retrieval-augmented assistant over official statute text from Pakistan Code or India Code where every answer cites specific sections, and a separate checker verifies that each cited section exists and that quoted text matches the source before the answer is shown. Measure hallucinated citations before and after the checker.
Learned LLM Router That Sends Easy Tasks to Cheap Models, with a Cost-Quality Frontier
Many LLM requests are easy enough for a small, cheap model, and a few need a strong, expensive one. Build a router that learns from your own evaluation data which requests need which model, compare it with simple rules and with always using one model, and plot the cost-quality frontier so a team can pick an operating point. Include cost tracking per request and a fallback when the cheap answer looks weak.
Sentence-Level Groundedness Checker for RAG Answers with Human-Calibrated Thresholds
RAG apps cite sources, but a citation does not prove the sentence next to it is supported. Build a drop-in checker that splits each answer into claims, checks every claim against the retrieved passages with a natural language inference model and an LLM judge, highlights unsupported sentences in the UI, and sets its thresholds from human labels rather than guesses. Test it on English and Urdu answers.
Single-Point Pressure Sensor for Household Water Leak and Tap Detection
Build a device that screws onto one tap in a home, samples water pressure hundreds of times a second, and works out which fixture is in use, how much water it used and whether there is a hidden leak. The research behind the idea was done on pressurised mains plumbing; most South Asian homes run on rooftop tanks with low gravity pressure. Finding out whether and how it works there is the core of the project.
Signed Web Page Archiving for Journalists and Fact-Checkers with Verifiable Captures
Build a self-hostable tool that captures a web page or social post exactly as it appeared, packages it in an open archive format, signs it and timestamps it, so anyone can later verify who captured it and when. Posts and pages are deleted or edited within hours, and screenshots are easy to fake. This gives newsrooms and fact-checkers evidence that holds up.
In-Browser Document Redaction for ID Cards, Bank Statements and Medical Reports
Build a web app that finds and removes personal details such as CNIC numbers, account numbers, phone numbers, addresses and names from PDFs and photos entirely inside the browser, using on-device OCR and small models, so the file never leaves the phone. People share copies of ID cards and bank statements with landlords, employers and agents every day, usually unredacted. The hard part is true redaction that removes hidden text, not black boxes drawn on top.
Wheat Crop Monitoring and District Yield Estimates from Sentinel-2 for Punjab
Build a pipeline that maps wheat fields across Punjab from free Sentinel-2 satellite images, tracks how the crop develops through the season and estimates district yields weeks before harvest. Wheat is Pakistan's staple, and procurement, import and flour price decisions depend on early estimates. The project tests how far open satellite data can go, with every estimate checked against published district statistics.
Checkout Dark Pattern Detector for South Asian Online Stores and Quick Commerce Sites
Build a crawler and classifier that walks real online store flows up to the payment step and detects deceptive design: items added to the basket without asking, fees that appear only at the end, fake countdown timers and shaming opt-out text, in English, Urdu and Roman Urdu. Earlier studies stopped at product pages. This project measures the checkout, where the money is, and gives consumer groups and store owners evidence they can act on.
Cloud Data Transfer Cost Attribution from Flow Logs for Small Teams
Build a tool that reads VPC flow logs and the cost and usage report from an AWS account, works out which service talks to which and over what path (same zone, cross-zone, NAT gateway, internet, cross-region), and attributes data transfer charges to the services and teams that cause them. Small teams see a large data transfer line on the bill and cannot tell where it comes from. The tool turns that line into a ranked list of flows with fixes.
Synthetic Patient Records Generator with a Privacy Attack Report for Hospital Data Sharing
Build a tool that takes a hospital table, generates synthetic records with several methods, and then attacks its own output to measure how much real patient information leaks. Hospitals want to share data with students and researchers, and synthetic data is often sold as automatically private. This project produces a report that shows both the usefulness and the privacy risk of each synthetic dataset before anyone shares it.
Signed, Staged OTA Updates with Automatic Rollback for ESP32 Device Fleets
Build a self-hostable update system for small fleets of ESP32 devices that signs every firmware image, rolls updates out in stages, checks device health after each update and rolls back automatically when something goes wrong. Small IoT companies in Pakistan and elsewhere ship devices with no safe way to fix bugs in the field, or with update schemes that a single bad build can brick. The project applies the security ideas behind automotive update frameworks at a scale a startup can run.
On-Device LLM Benchmark App for Budget Android Phones with Urdu Quality Tests
Build an Android app that runs small language models on the phone itself and measures what developers need to know before shipping on-device AI: speed, memory, battery drain, heating over a sustained run, crashes, and answer quality on Urdu and Roman Urdu tasks. Results from volunteers' phones feed a public table of how common budget models actually perform. Most published numbers come from flagship phones and English prompts, which tells a developer in Pakistan very little.
Offline Urdu Voice Commands on a Microcontroller for Home Lights and Fans
Build a wall-mounted controller that recognises a small set of spoken Urdu and Punjabi commands, such as turning a fan or light on and off, entirely on a microcontroller with no internet and no cloud. Older people and people with limited mobility often cannot reach switches easily, and cloud voice assistants need connectivity and understand Urdu poorly. The core is a tiny neural network trained on a dataset you record yourselves.
Spreadsheet Formula Error Finder for Small Business Workbooks
Build a static analyser for Excel workbooks that finds formulas which break the pattern of their neighbours, ranges that miss a row, hard-coded numbers inside formulas and references to empty cells. Small businesses run pricing, payroll and stock on spreadsheets that nobody audits. The tool explains each suspicious cell in plain language and is evaluated against a labelled set of real workbook errors.
On-Device Language Model NPCs with Memory and Lore Consistency Checks for Indie Games
Build a small framework that lets non-player characters in an indie game hold free conversations using a small language model running on the player's own machine, with memory of past events and a checker that stops them contradicting the game's lore or breaking quests. Cloud-based character services charge per conversation and need a connection. The core of the project is measuring and reducing contradictions, not just making NPCs talk.
Crop Fire Smoke Attribution for Lahore Smog Using Satellite Fire Data and Wind Back-Trajectories
Build a daily analysis that estimates how much of Lahore's smog on a given day is linked to crop residue fires upwind, in Pakistani and Indian Punjab, and how much is local. It combines satellite fire detections, wind back-trajectories and PM2.5 readings. Every winter the cause of smog is argued about; this project replaces the argument with a transparent, repeatable daily estimate.
Helm Chart Permission Analyser That Finds Kubernetes Cluster Takeover Paths Before Install
Build a tool that reads a Helm chart before it is installed and works out whether the permissions it grants could let an attacker who compromises one of its pods take over the whole cluster. Many third-party operators and add-ons ask for far more access than they need. The analyser explains each dangerous permission, the attack path it opens and a narrower alternative.
Browse project ideas by domain and difficulty
Each of these is its own page, so you can send someone straight to the AI ideas or the beginner ones.
Stuck on something specific?
Writing only gets you so far. If you want an answer to your situation rather than the general case, book a session and we will work through it together. Sessions are free for approved Sefism members, and a few slots open each week.
- 1:1 Mentorship
- Fresh Graduate Mentorship
- Career Guidance
Brand deals, promos and partnerships: tauseeffayyaz@gmail.com
Follow along
New writing, resources and project ideas land here first.