Machine Learning Based Natural Language Querying of Automotive Data

This research focuses on developing a retrieval system for JATO Dynamics in the field of automation. With recent advancements in artificial intelligence, user expectations have significantly increased—particularly the demand for agile, efficient, and user-friendly access to relevant data. However, not all users are proficient in formulating precise queries or prompts to retrieve the information they need. This highlights the need for a system capable of understanding natural language input, enabling accurate retrieval even from approximate or semantically similar queries. While many such applications rely on Large Language Model (LLM) APIs, these solutions pose risks such as potential data leakage and high operational costs, especially in use cases requiring frequent API calls. Therefore, the goal of this study is to develop an in-house retrieval system that leverages syntactic and semantic techniques—through Machine Learning or a hybrid approach—to provide secure, efficient, and cost-effective natural language-based search capabilities.