Advanced Machine translations using agentic LLMs

Since Dometic operates on a global scale, having a unified language in how we present and talk about our products is essential. Our brand identity should remain consistent, no matter the language. Marketing materials, sales documents, and product descriptions all need to be tailored to their specific audiences and localized appropriately, while still preserving the original tone of voice.
The latest advancements in NLP and Generative AI offer a swift, interactive way to achieve this. While traditional machine translations have been useful for straightforward translations, using LLMs in an agentic manner can provide more sophisticated solutions. Additionally, they can serve as an excellent tool for sparking ideas among copywriters, boosting their productivity.
This Master Thesis will investigate the possibility of generating better machine translations that keeps the tone of text and /or domain technical terminology when translating different types of texts by using state of the art generative AI setups. This includes investigating agentic setups and logic as well the need for finetuning an own model. The project will be conducted from Dometic’s Headquarters in Solna and preferably by two persons as we are looking for a thorough investigation of the subject. You will get a chance to work with a big global company within the manufacturing industry and encounter many different applications for the technical solution.

Keywords: Machine Translations, NLP, Generative AI, Agentic AI, LLM