Edge-Based AI Manager for Real-Time Business Analytics and Interactive User Assistance

The project involves developing a robust AI manager agent designed to operate entirely on edge computing devices, specifically targeting small to medium-sized enterprises (SMEs) in sectors like hospitality. The AI manager will collect and analyze real-time data from various applications used by SMEs—such as order management, inventory control, customer feedback, and more—to provide actionable insights, forecasts, and recommendations to improve business operations. A core component of the project is creating an interactive, conversational AI interface that allows business owners to ask questions in natural language (text or voice). The AI manager will respond by summarizing current business performance, identifying key metrics, and providing step-by-step guidance for using business applications. An essential feature of this system will be its ability to self-learn from data trends and user feedback, making it an adaptive and evolving tool tailored to each business. Students will address several challenges, including developing lightweight, efficient AI models optimized for edge processing, creating a user-friendly dashboard with visual guidance, and ensuring the AI's modular design to support integration with third-party applications. Additionally, the project will require maintaining strict data privacy and security due to the sensitive nature of business data.

Automated Edge Deployment Platform: Frontend and IAAS Integration for Distributed Application Management

This project aims to empower customers to deploy, manage, and monitor their applications directly on our custom IAAS edge infrastructure via the Shiriki Cloud frontend. The Shiriki Cloud platform will provide users with a seamless, automated experience for deploying containerized applications on distributed Balena OS-based edge devices. The IAAS infrastructure will enable resilient and autonomous edge deployment by incorporating lightweight Kubernetes (K3s), KubeEdge for edge autonomy, Ceph for distributed storage, and Cilium for secure, policy-driven networking. The project’s challenges include orchestrating large-scale application deployments across geographically dispersed edge devices, ensuring data resilience, and integrating secure communication. The frontend must provide a user-friendly interface that abstracts the complexities of edge infrastructure, allowing customers to deploy applications at scale without requiring in-depth technical knowledge.

Modular AI for Real-Time Video Analytics on Edge Devices

The objective is to develop a single, modular AI model for edge devices that can perform multiple real-time video analytics tasks, including customer flow analysis, incident detection, security monitoring, and compliance tracking, while being optimized for edge hardware and ensuring GDPR compliance. Tasks and Activities: Model Development : Build a shared backbone AI model with task-specific outputs for modular functionalities. Optimize the model for edge devices by implementing quantization and pruning techniques. Edge Integration : Develop containerized modules for dynamic task activation on Balena OS. Implement real-time processing for video analytics tasks like object detection, tracking, and incident alerts. GDPR Compliance : Integrate face-blurring and anonymization features into the model for privacy protection. Performance Testing and Optimization : Test and optimize the model across various edge hardware scenarios (e.g., single or multi-camera setups). Ensure the system supports OTA updates for easy deployment and maintenance. Deliverables: A fully integrated, modular AI model capable of performing multiple tasks on edge devices. A containerized system for easy deployment, management, and updates via Balena OS. A GDPR-compliant, real-time video processing system with dynamic task activation and resource allocation.

EdgeApp Hub: Decentralized App Store for Seamless Offline & Online Access

To develop a decentralized app store platform, EdgeApp Hub , that enables users to download, update, and manage Android applications both offline and online through distributed edge nodes. The platform will leverage P2P communication, AI-driven app recommendations, and secure infrastructure for a scalable and user-friendly solution. Tasks and Activities System Architecture Design Define the technical architecture, including edge nodes, P2P protocols, distributed app repositories, and AI modules. Design data flows for app downloads, updates, and caching. Backend Development Build the backend system for managing app metadata, versioning, and P2P communication. Implement distributed databases or ledgers for app indexing and metadata storage. AI Integration Develop AI models for app recommendations, user behavior analysis, and predictive caching. Optimize models for resource-constrained edge devices. Frontend Development Create a user-friendly web/mobile app store interface with offline and online functionalities. Include features like app search, recommendations, and update notifications. Security Implementation Ensure app validation with cryptographic signatures and integrity checks. Implement end-to-end encryption and secure user authentication. Testing and Optimization Conduct functional and performance testing on edge nodes and P2P networks. Optimize the platform for low-bandwidth environments and seamless offline access. Deliverables Technical Design Document Comprehensive documentation of the architecture, system workflows, and protocols. Functional Prototypes A working prototype of the app store interface with core features. Operational P2P communication between edge nodes. AI Recommendation System A trained and deployed AI module for app recommendations and caching. Deployment Package Fully containerized application stack ready for deployment on edge nodes. User manual and installation guides. Testing Report Detailed report on testing results, including performance, security, and user feedback. This project will provide learners with hands-on experience in edge computing, decentralized systems, and AI integration, preparing them for real-world challenges in tech innovation.