Introduction

November 19, 2025, was the start of a rollercoaster week.

Our team, NeuroHack, representing the University of Sharjah, participated in the function1 AI Genesis Hackathon at Festival Arena, Dubai. It was genuinely one of the most incredible experiences I have ever had.

The scale of the event was massive, with 4,842 participants and 605 teams from around the world.

Out of these, around 300 teams were invited for the onsite hackathon, and we were proud to be one of them.

Innovation

We developed PRISM (Personal RAG Integrated Semantic Multimodal) specifically for the Qdrant Challenge track. The project addresses the fragmentation between code, documents, and images.

By leveraging Qdrant's advanced vector search engine as the core memory layer, we created a system where users can query multimodal data by using natural language. This marks a new era in education.

What we built

We implemented several features to improve the research and study workflow, including:

• Vector Insights Graph to visualize content relationships using Qdrant embeddings

• RAG chat interface powered by Google Gemini for context-aware answers

• Smart recommendations to discover related content through semantic similarity

Together, these features aim to boost productivity and make learning, research, and knowledge discovery more intuitive.

Participating in this challenge also gave us the opportunity to dive deeper into 768 dimensional vector embeddings, multimodal AI agents, and the power of semantic search.

Challenge

1. During the hackathon, we faced an unexpected obstacle: a massive global Cloudflare outage but that did not stop us.

2. Despite the challenges, we stayed focused, adapted quickly, completed our build, and successfully delivered our solution on time.

3. It was a powerful lesson in resilience, teamwork, and building under pressure.