Synthetic catalogue
50 products across 15 categories with random brand/style/price.
Live · Retail & CPG
Recommendation engine
Pick a customer persona on the left. The model returns the top recommended products from a synthetic catalogue of 50 items, computed via cosine similarity on a user-item interaction matrix.
Pick a customer
Recent purchases
- No prior purchases.
Recommended next
How it works
Synthetic users
6 personas each with a category preference distribution drawn from the catalogue.
Interaction matrix
A 6 × 50 binary matrix indicating which user purchased which product.
Item-item similarity
Cosine similarity computed on the transpose of the interaction matrix.
Score & rank
For a target user, score = Σ similarity × prior purchases. Already-purchased items are hidden.
Production swap
Same architecture works at scale with LightFM, implicit, or a transformer-based recommender — same API, same UI.
Want this on your real catalogue?
We build recommendation systems with your inventory, real interaction logs, business constraints (margin, cold-start, fairness), and integration into your existing storefront or app.
Ready to start
Turn one AI use case into measurable production value.
Book a 30-minute consultation. We will walk through the use case, sketch the value case, and tell you honestly whether we can help.