Case study / 003Final year projectFull-stack · RAG · SaaS
LAWVERARAG SAAS.
Role / Full-Stack Engineer
A web-based SaaS platform exploring how retrieval-augmented generation can make legal information workflows more focused, traceable, and useful.
Conceptual SaaS product viewPortfolio visualization / Web application
Problem
Legal information is only useful when the right context can be found.
The project explored a focused product problem: how a web application could help a user ask a legal-information question, retrieve relevant source material, and receive a response grounded in that retrieved context instead of relying on an unconstrained model answer.
Product approach
RAG designed as a workflow, not a chat box.
Lawvera was shaped as a SaaS web application with explicit input, retrieval, context, response, history, and administration surfaces. The interface needed to make the journey understandable while the system coordinated documents, search, model interaction, and saved results behind it.
Engineering scope
Ownership across product, data, and AI integration.
Responsive SaaS web application and authenticated product flows
Document ingestion and retrieval-oriented data workflow
Backend coordination around retrieval and generated responses
Usable loading, empty, success, and failure states
Administrative visibility for managing the product workflow
Full-stack delivery carried as a final-year engineering project
Outcome
A complete SaaS concept built around responsible context.
The result demonstrated more than a model integration. It connected retrieval, generated output, product states, and web application structure into a coherent platform concept—and strengthened my practical understanding of how RAG belongs inside a full product system.
Project note
Built as an academic product, presented as engineering evidence.
Lawvera was my final-year project and a web application, not a mobile app. The portfolio visual is a conceptual showcase created to communicate the product category and workflow clearly; it does not represent client data or a third-party production interface.