SIT Portfolio

CS64-RE03 :: Mari: Automatic response to questions related to the legal debt collection act of 2015 using artificial intelligence and natural language processing techniques on a website platform

มะลิ: ระบบตอบคำถามอัตโนมัติเกี่ยวกับการทวงถามหนี้ตามพระราชบัญญัติทวงถามหนี้ พ.ศ. 2558 บนเว็บไซต์โดยใช้เทคโนโลยีปัญญาประดิษฐ์และประยุกต์ใช้เทคนิคการประมวลผลภาษาธรรมชาติ

details

Mari is a system designed to provide legal consultation services related to the Thai Legal Debt Collection Act of 2015 in Thailand. Utilizing artificial intelligence (AI) and natural language processing (NLP) using the retrieval-augmented generation (RAG) techniques based on the SambaNova pre-trained model. The service will be available on the website, requiring users to input queries in Thai, with responses provided in Thai as well. Mari aims to empower individuals, both creditors and debtors, with essential legal knowledge and guidance. By offering an accessible and user-friendly platform, Mari aims to reduce the time and expenses associated with legal consultant and improve understanding of the legal implications of debt collection decisions.

tools & techniques

Frontend development tools

- Programming language: Typescript

- Design framework: MaterialUI

- Web framework: React

- HTTP Client: Axios Axios

Backend development tools

- Programming language: Go

, Python

- Framework: GoFiber GoFiber

, FastAPI

- OpenAPI Specification: Swagger

Database management system

- MongoDB

: Document database suitable for storing conversation history or any other unstructured data. - Chroma

: Vector database for storing embeddings that simplifies building LLM apps. It offers advanced querying, filtering, and scalability, with support for Python notebooks.