Systems and methods for pre-processing domain-specific translation models

A computer-implemented method, comprising:

• receiving an inventory string associated with at least one product within a specified domain;
• generating, via a generative artificial intelligence (GAI) model, a plurality of attributes based on the inventory string;
• applying one or more deterministic filters to a plurality of product types and the plurality of attributes to identify at least one candidate product type from the plurality of product types;
• and
• transmitting at least one candidate product type to a translation model trained on the specified domain and configured to identify the at least one product.

A non-transitory computer-readable medium embodying a program that, when executed by at least one computing device, cause the at least one computing device to:

• receive a request comprising an inventory string;
• generate, via a generative artificial intelligence GAI model, a plurality of attributes based on the inventory string;
• apply one or more deterministic filters to a plurality of product types and the plurality of attributes to identify at least one candidate product type from the plurality of product types;
• applying a translation model to the at least one candidate product type and the plurality of attributes to identify a requested product, wherein the translation model is trained on a plurality of synthetic input-output pairs, each of the plurality of synthetic input-output pairs translate one or more naming variations to the plurality of product types;
• determine that the requested product is in stock based on an inventory database;
• and
• transmitting the requested product in response to the request.