Show HN: I built a RAG engine to search Singaporean laws (github.com)

The Singapore Intelligence RAG System has been launched as an innovative platform leveraging AI to provide precise information on the legal framework, policies, and historical context of Singapore. This system distinguishes itself from traditional language models by employing Retrieval-Augmented Generation (RAG), utilizing a meticulously curated dataset of over 33,000 pages of relevant Singaporean documents to minimize the instances of "hallucination" common in many AI systems. Significantly, the RAG engine is optimized for low-resource environments, utilizing a high-performance pipeline that includes BGE-M3 for semantic embeddings and FAISS for rapid vector retrieval. With a robust "Triple-Failover" architecture ensuring 99.9% uptime and a locally-executed embedding model enhancing privacy and reducing latency, this platform sets a new standard for reliability in AI applications. The frontend, built with React and Framer Motion, creates an interactive user experience while avoiding API calls for vectorization, ensuring both efficiency and security in handling sensitive legal data. This development has vast implications for the legal tech landscape, potentially transforming how legal research is conducted in Singapore and beyond.