Hi Sir, I can create a search engine using python (of course it will be an api). The search for a similar image between 1000 images will take 10 milliseconds. so search within 10,000,000 images could take up to 10 seconds I'm afraid. But we can make the search asynchronous, once a matching result is found, it will be sent to the api client (browser/mobile) right away and the client will keep pinging the endpoint until the search is complete by assigning a unique ID for the search. Can be made more faster by generating a file with a cache of image feature extraction data so that we can refer that file instead of reading all the images, thus will make the process very fast. This file generation can be done using a time where the least amount of your site visitors are active (sleeping times). If you understand the risks and that the algorithms can have both accurate and invalid results as well, then I'm ready to take this project.