You work at a marketplace with 400 million product listings, each with 1 to 8 photos. Sellers write inconsistent titles: "vtg denim jkt L" is a real listing, so keyword search fails on a large fraction of the catalogue.
Product wants two capabilities:
The catalogue turns over quickly: 4 million new listings a day, and roughly the same number sold or delisted. Query volume is 8,000 searches per second at peak.
Build the architecture on a canvas: place the components, configure them, connect them into a data flow, and write a short reason for each one. The AI reviewer grades your design against a rubric written specifically for this problem.
Work out the memory for 1.2B vectors at 512 dimensions, then show me how you get it under budget.
A new listing must be searchable in 15 minutes but your index rebuild takes 6 hours. How?
A buyer filters to 'available in Germany under 50 euros' and gets 2 results from a top-200 retrieval. What went wrong and how do you fix it?
Minimum 6 components · needs a wide desktop screen