Machine Learning for Product Matching

We have 2 databases. 1 - Products in Magento, separated by attributes (11,000). 2 - Warehouses (23,000) - sku, description, stock, price, warehouse. We want to match one database with the other. However, the names of the products are not always identical. For example: Magento product: Cooper Discoverer AT3 235/75R15 104/101R Warehouses: Cooper AT-3 235 75 r15 104R , 4pcs, $867 Discoverer Cooper AT / 3 235/75R15 101R, 12pcs, $842 Rim 235/75 R15 Cooper Discov A-T3 235-75-R15 104/101 R, 185pcs, $452 104 / 101R Cooper Discoverer AT3 235/75r15 on sale, 785 pcs, $459 These products are the same, but the names of their products vary a lot. Functions: - Use of synonyms and stopwords: “usa, unites states, u s a, united states of america”. (admin panel) - Rules to select the best stock and price in stores (put the best price if it is greater than 12 pieces, if not put the highest). (admin panel) - Compare price vs competition and determine price. (add another competency database) - Additional rules (brand "Hankook" + 30%, Size "R15" -8%, etc.) (admin panel) - Update product prices in Magento. (api) - Ultra Fast and process viewer with percentage. (admin panel) Preference: Phyton, JSON, PHP, MySQL, elasticsearch, Software Architecture. Possibility to start the project immediately.