Ali Soban@ali-soban

Smartstudy is the next student companion! It is an AI-powered application, which has the following capabilities: - Summarise content - Generate questions and answers from given content - Generate quizzes - Explain and debug code The application utilises the Granite model to provide these services. Tuning to specific prompts and datasets has allowed better and relevant performance, along with a robust UI, where Python has been utilised to give the user a simple but rich experience while accessing the application. In order to connect the application to relevant models, Python has been utilised for data handling and transfer. The ability of the application to read different files, as well as code debugging and generation make it a complete package, with options of quiz generation making it a complete preparation package. Future goals with respect to this application include adding features that enhance user experience, such as text-to-speech generation, multi lingual capabilities and diagram generation to name a few.

The application utilises the Llama-3.3-70b-versatile model for analysis of the upload and download speeds, along with latency and bandwidth of a given region, and predicting the development score based on these parameters. The model further builds on it by suggesting improvements based on priority and condition of network in that region, while also providing detailed steps on how to achieve the desired target. RAG was implemented in order to enhance the performance of the model, which added great value in the performance of the model and helped us achieve much accurate and relevant results. CSV file with over 13000 unique instances of data was utilised for tuning the model for accurate results.