← All projectsAI / Machine Learning · Python
Recommendation System — hybrid collaborative + content filtering
Recommends from real user behaviour, explains why, and handles the new-user problem.
Most recommender projects hardcode a similarity matrix, demo on the same three users, and freeze the moment an examiner logs in as someone new. This one builds collaborative filtering with matrix factorization, adds a content-based model for cold-start, blends the two, and shows why each item was recommended. It reports precision and recall honestly rather than claiming it is always right, and it answers the question that sinks most of these: what do you recommend to a user with no history?
- Level
- Advanced
- Live sessions
- 8
- Seats per college
- 1 / year
What it does
- Personalised recommendations from a user's ratings and behaviour
- Content-based fallback so a brand-new user still gets sensible picks
- A plain-English 'why this was recommended' next to each item
- Honest evaluation — precision@k and recall, not a vanity accuracy
- Admin view of the catalogue and the interaction data
- A FastAPI service the React front end calls for live recommendations
What the sessions cover
- 01Collaborative vs content-based vs hybrid — and when each fails
- 02Preparing the data and building the user-item ratings matrix
- 03Matrix factorization (SVD) — the intuition without the heavy math
- 04Content features and similarity for the cold-start path
- 05Blending the two models and handling a brand-new user
- 06Evaluating honestly: precision@k, recall, and why accuracy lies here
- 07Serving with FastAPI and rendering recommendations in React
- 08Deployment and a full mock viva on the cold-start design
Stack
Pythonpandasscikit-learnSurpriseFastAPIReactPostgreSQL
Pick what you need
Direct UPI / bank transfer in monthly installments. Sessions start after the first installment; handover is after the final installment. Full terms — including what happens if you stop midway.
Source
₹6,000
single payment
- Full source code
- Local setup guide
- Sample dataset / seed data
- One handover call
Academic
₹11,000
or ₹5,500/month × 2
- Everything in Source
- Project report (synopsis + full document)
- Presentation deck
- ER / DFD / architecture diagrams
- Deployment walkthrough
Mentored
Most picked₹18,000
or ₹6,000/month × 3
- Everything in Academic
- 8 live sessions on Google Meet
- Line-by-line code walkthrough
- Viva question prep + mock viva
- One module varied for you (dataset / domain)
- WhatsApp support until submission
Opening 3 seats
Claim a seat on WhatsAppThis page has no reviews on it yet. Be the reason it does.
The first 3 students get 40% off in exchange for an honest review and permission to show your finished project here. Same build, same sessions, same handover — the discount is the price of going first, and it is not coming back once those seats are gone.
You are about to send money to a stranger on the internet.
Fair. So here is everything you need to check who I am — all of it public, all of it older than this page.
Who builds itSenior Full-Stack AI EngineerHyderabad, India. Full background, work history and CV.Check it →Shipped, not claimed4 productsBuilt solo and deployed. Architecture teardowns for each one.Check it →Code in the openPublic reposThe same hands that write your project. Read it before you pay.Check it ↗Professional recordLinkedInVerifiable employment history, not an anonymous seller account.Check it ↗
Two things I hold myself to
- Only 1 students from your college get this project in an academic year. On the Mentored tier one module is varied for you — the dataset, the domain, or a feature — so no two submissions read the same.
- You build nothing. You just have to be able to explain it. The project is delivered finished — code, report, diagrams, deployed. Then daily 60-minute sessions on Google Meet after 8:00 PM IST, walking the code you are submitting, ending in a mock viva. Turn up and you will know it well enough to answer anything the panel asks.
Enquire about this project
Tell me your submission deadline and which tier fits. You get a session schedule and an installment split back — usually within a day.
Questions students ask
- Do I have to build any of it myself?
- No. The project is built, documented, and deployed for you — code, report, deck, diagrams, and the running application. Your only job is to turn up to the sessions and learn what you are submitting.
- Can I get my own idea built instead of picking from the list?
- Yes. Send whatever your guide approved — or an idea you want built from scratch — through the custom build form. Any stack, any domain. Custom projects are quoted between ₹10,000 and ₹50,000 depending on scope, and run on the same sessions and installments.
- Do I get the source code, or only the explanation?
- You get the full source code on every tier. The Mentored tier adds live sessions where the code is walked through line by line so you can defend it in your viva.
- How do the daily sessions work?
- Sessions are live on Google Meet, one per day, starting after 8:00 PM IST so they do not clash with college hours. Each runs about 60 minutes and covers one module of the project.
- Can I pay in installments?
- Yes. Payment is direct — direct upi / bank transfer — split into monthly installments. Sessions start after the first installment; the complete source, report, and deployment handover comes after the final one.
- I can buy a project online for ₹2,000. Why does this cost more?
- Because that ₹2,000 buys a zip file. Nobody explains it to you, the same file has been sold to hundreds of students, and it collapses the moment your examiner asks why you chose that database. What you are paying for here is a project built for you, capped per college, and the sessions that make you able to answer for it. If a zip file is all you need, buy the zip file.
- How do I know you will not disappear after I take money?
- Payment is split across months, so you are never paying for work you have not seen — you watch the code in sessions from the first one. If I fail to deliver what the project page promised, your remaining installments are cancelled and the last one is returned. I also have a name attached to this: the same site carries my portfolio, my published work, and my public repositories. That is not something I would burn for one project.
- What if my guide rejects the project or asks for changes?
- Changes your guide asks for during review are made at no extra cost — that is part of the build, not a new scope. If your guide rejects the topic outright before work begins, you can switch to a different project or stop with nothing further owed.
- Will another student at my college submit the same project?
- No. Each listing is capped to a limited number of students per college per academic year, and the Mentored tier varies one module — the dataset, the domain, or a feature — so two builds are never identical.
- Who actually builds and teaches this?
- Yaseen Khatib — a Senior Full-Stack AI Engineer. The same person writes the code and runs the sessions, so every answer in your viva prep comes from whoever wrote the line you are asked about.