During my time at Cardinal LLC, I led an R&D project aimed at creating a recommendation engine that analyzes and profiles user data regarding past searches on the Freename.io web3 domain registrar platform. The developed system is able to intelligently provide personalized and relevant recommendations for similar search terms the user might be interested in. I opted for a combination of open-source and commercially available generative models such as Google’s Gemma, Meta’s Llama 3, Mistral AI’s Mixtral, as well as OpenAI’s GPT 3.5/4/4o, Google’s Gemini 1.0/1.5 and Anthropic’s Claude that can be seamlessly switched from one to the other using the developed API. The final phase focused on the real-time implementation and testing of the recommendation engine using FastAPI in a Dockerized environment, harnessing the few-shot learning technique to provide accurate recommendations for each user. Throughout the project, I conducted an in-depth analysis of both open-source and commercial Generative AI models which were ultimately included in the final report for the project, along with examples of generated outputs that significantly enhanced the user experience and service optimization.