High Speed Machining (HSM) is a critical process notably found in precision machining, where selecting optimal tools can significantly improve performance. Traditional methods for selecting the optimal tool rely heavily on operator experience, leaving room for automatic solutions. To address this challenge, we propose an innovative recommendation system that leverages machine learning techniques through Deep Factorization Machines (DeepFM). The proposed method utilizes a dataset comprised entirely of theoretically derived annotations provided exclusively by experts in the field of tool manufacturing. Furthermore, we incorporate the geometric characteristics of each tool extracted from a database following the ISO 13399 technical standard. We demonstrate that this implementation is reliable and propose a detailed user feedback framework that will be used to tune the dataset through lifelong learning.