CINI IT Challenge
Tackling real-world problems with modern technologies in a university-level competition.
Event Context
The CINI IT Challenge is a national competition for university students, organized by the National Interuniversity Consortium for Computer Science. The goal is to solve real-world problems with innovative software solutions.
Participating meant competing with students from all over Italy, working under pressure and presenting the project in front of a panel of experts.
Overview
Project developed for the CINI IT Challenge (Smart Cities & Communities), organized by the National Interuniversity Consortium for Computer Science.
Mobishare is a smart mobility platform for sharing bikes and scooters, with a live map updated via MQTT and an advanced AI agent (OllamaSharp) able to answer users and automate tasks through RAG and function calling.
Challenge
The exam requirements constrained the domain and core technologies. The main challenges were the secure, real-time management of vehicle GPS data via MQTT and the integration of a truly useful and autonomous AI agent.
Moreover, integrating real-time architectures was quite a hurdle to overcome.
- MQTT for real-time vehicle location: GPS data published from Arduino and instant map updates.
- AI agent with OllamaSharp: RAG system and function calling to answer users and automate tasks.
- Security and architectural complexity: multi-container Docker orchestration, secrets management, authentication, and payments.
Technical Deep-Dive
Integrating MQTT, SignalR, and OllamaSharp required careful design to ensure performance and security. I orchestrated Docker containers to isolate services and manage secrets securely.
The AI agent leverages RAG and function calling to answer complex questions and automate tasks, while the live map updated via MQTT provides a truly interactive user experience.
Solution
Modern stack: ASP.NET Core backend, Razor Pages frontend, SignalR for real-time, Docker Compose, SQLite, Google Maps SDK, PayPal SDK, Google OAuth.
MQTT and real-time location. The system receives vehicle positions in real time via MQTT, ensuring instant and reliable updates.
Autonomous AI agent. The AI agent, integrated with OllamaSharp, autonomously executes queries and tasks thanks to RAG and function calling, improving the user experience.
Security and interoperability. Secure data management, Google authentication, digital payments, and Docker orchestration for a robust infrastructure.
Related
A local-first, privacy AI second brain. Multi-agent RAG turns chaotic Markdown vaults into a queryable archive — and a navigable 3D galaxy of ideas.
Artificial Intelligence didn’t start with ChatGPT. From its origins in the 1950s to today’s deep learning models, let’s explore what AI really is, how it learns from data, and why it’s not just computer “magic.”
Do you have a complex architecture to build?
Tell me about your idea. Let's design a solid, automated, and scalable solution together.