From mid-April to late June 2024, I had the privilege of completing an enriching internship at Robert Gordon University (RGU) in Aberdeen, Scotland.
This experience allowed me to contribute to an innovative project while developing advanced technical skills in an international environment.
An Ideal Learning Environment
My internship took place within the IT department, located in the Sir Ian Wood Building—a modern space dedicated to research and technological innovation.
Working in this setting enabled me to collaborate with international researchers and students, fostering constructive exchanges and a culturally stimulating scientific experience.
Developing an AI-Driven Web Application
The main objective of my internship was to design and develop an interactive web application that integrated cutting-edge technologies:
- Frontend: Built a dynamic and user-friendly interface using ReactJS.
- Backend: Developed efficient APIs with FastAPI.
- Data Management: Implemented real-time databases and secure authentication systems with Firebase.
- Artificial Intelligence: Integrated an Explainable AI (XAI) model to analyze loan applications and provide actionable solutions to users.
Focus on the XAI Model for Loan Applications
One of the most innovative aspects of the project was the use of an XAI (eXplainable Artificial Intelligence) model.
This model aimed to process loan applications transparently and comprehensibly through several key steps:
-
User Data Analysis The model analyzed parameters such as income, credit score, age, and the requested loan amount.
-
Decision-Making Based on training data, the model assessed whether the loan application could be approved or denied.
-
Decision Explanation In case of rejection, the model provided clear explanations, identifying which criteria were blocking the approval.
-
Solution Proposals Using the DiCE library (Diverse Counterfactual Explanations), the model generated actionable changes to turn a rejected application into an approved one.
For example, it could suggest increasing income or reducing the loan amount.
This approach made the AI system more accessible and user-friendly by offering practical guidance to users on how to improve their applications.
Challenges and Achievements
The main challenges of the project were creating the XAI model and integrating it into the web application.
These tasks required a deep understanding of explainable AI concepts and learning to use the DiCE library effectively.
With the guidance of my supervisor, Pedram Salimi, and through extensive self-research, I successfully overcame these hurdles.
Results and Future Perspectives
The final application is both ergonomic and powerful, capable of analyzing complex scenarios and providing users with clear and actionable feedback.
This project significantly strengthened my expertise in web development and AI while confirming my interest in pursuing a career in these fields.
In conclusion, my internship at RGU Aberdeen was a defining experience, combining technological innovation with cultural learning.
It allowed me to contribute to a project that has a tangible impact on how users interact with intelligent systems.
The full internship report is available here.