A line which moves indefinitely
Bishop & Co. logo

International Mentoring Centre

The International Mentoring Centre was established at the University of Szeged to provide essential guidance through the Erasmus Student Network Szeged, assisting students with various aspects of their everyday lives. The University of Szeged (SZTE), a prestigious higher education institution in Hungary, is renowned for its commitment to quality education across its twelve faculties. Ranked as the best Hungarian university by ARWU in 2024, it boasts a rich tradition and a history of notable professors, including Nobel Laureates Katalin Karikó (2023) and Albert Szent-Györgyi (1937).

Open WebsiteAn arrow going through a rectangle&s top right corner
A grid of dots
A grid of dots
AI Chatbot Assistant made for University of Szeged

Challenges

Implementing advanced AI technologies

The growing international interest SZTE faces led to a surge in inquiries to the International Mentoring Centre. They required a sophisticated AI solution to aid the recruitment of international students, that provides accurate and contextually relevant responses to users. We decided to develop a Retrieval Augmented Generation (RAG) system that integrates industry-leading AI technologies while maintaining lightweight architecture and quick development cycles. The solution needed to be scalable, cost-effective, and easily manageable.

Multiple information source

Since the information provided by the International Mentoring Centre is spread across multiple web pages, it’s harder for prospective students to find details on study programs, student visas, and accommodations. Our solution had to ensure that they could see everything in one place. It also needed a flexible infrastructure that integrates seamlessly with existing systems.

Maximizing cost efficiency

Another significant challenge was to develop a high-performance system, that stays flexible and proves to be easy to update and maintain, without incurring substantial costs. We had to leverage serverless technologies and optimized resource usage, to support our partner’s passion for continuous innovation. Additionally, the deployment process had to be streamlined for continuous delivery and quick iterations without compromising quality.

Our Solutions

Custom AI chatbot with RAG System

We architected and developed a custom AI chatbot by creating a Retrieval Augmented Generation (RAG) system. This approach combines the strengths of information retrieval and natural language generation to provide accurate and contextually relevant responses. This framework ensures the AI won’t fabricate information and avoids hallucinations. Utilizing the LLama3 Large Language Model (LLM) ensured the chatbot could understand and generate human-like text based on the retrieved information.

State-of-the-art cloud infrastructure

To support the advanced AI functionalities, we implemented the solution using AWS Bedrock and Pinecone. AWS Bedrock provided a robust foundation for building and managing the machine learning models, while Pinecone facilitated efficient vector search and retrieval. The entire infrastructure was managed using AWS Cloud Development Kit (CDK), enabling infrastructure as code and ensuring reproducibility and scalability.

Serverless and cost-effective deployment

Our solution leveraged AWS Lambdas and other serverless technologies, ensuring the system was lightweight and cost-effective. By adopting an AWS-native approach, we achieved near-zero operational costs while benefiting from the scalability and reliability of AWS services. The serverless architecture allowed us to handle varying loads efficiently without constant infrastructure management.

A grid of dots
A grid of dots
Phone, and tablet view of the RAG Ai assistant chatbot made for University of Szeged

Results

Rapid development and deployment

The advanced AWS services and serverless technologies enabled us to rapidly develop and deploy the custom AI chatbot, streamlining the development process for quick iterations and continuous delivery of new features. This agility ensured the solution could adapt to evolving requirements promptly, resulting in an AI chatbot that drastically reduced the manual workload of our partner’s dedicated team. The chatbot handles the ever-growing number of repetitive questions from incoming students and reduces the need for a tedious update process across multiple sites. This allows the International Mentoring Centre to focus on other priorities and provide more effective support for complex student inquiries.

Scalable single source of truth

The RAG system, powered by LLama3 and supported by AWS Bedrock and Pinecone, delivered high performance in generating accurate and contextually relevant responses. The serverless architecture provided automatic scaling capabilities, ensuring the system could handle increased loads without compromising performance. Thanks to our solution students only need to use this page to receive factual, and trustworthy answers to all their questions about their future studies in Szeged.

Cost-effective system update

The International Mentoring Centre of the University of Szeged is the first in Hungary to revolutionize communication with students through new technologies such as AI. We were delighted to collaborate with them and develop an innovative solution. Thanks to the serverless approach and efficient resource management, the operational costs of the solution were kept to a minimum. The AWS-native solution ensured that we only paid for the resources consumed, resulting in significant cost savings while maintaining high availability and reliability.

Testimonial

Thanks to Bishop & Co. and their understanding of AI solutions, the Erasmus Student Network successfully implemented generative AI into their workflows from scratch in just two months. This solution has significantly reduced repetitive tasks and manual work hours, allowing our team to focus more on building personal relationships with new students. Their speed, excellence, and commitment to quality make them an outstanding partner.

Dr. Tamás Bene
Dr. Tamás Bene Director for International Affairs and Public Relations University of Szeged

Do you have an ongoing project or would you like to start one?

Contact us and share your vision. We are here to help you.

A line which moves indefinitely

More Case Studies

SZIN mobile application displayed on a Google Pixel and an iPhone

SZINPictogram of a link

Hungary's longest running festival in Szeged, attended by tens of thousands each year.

Mobile App
Cross-Platform