Manual quality inspection can be slow, inconsistent, and difficult to scale. We developed an AI-powered visual inspection system that identifies screw types and detects manufacturing defects in real time. Optimized for resource-constrained edge hardware, the solution delivers rapid predictions and gives operators a simple interface for monitoring inspection results.

We developed a computer vision system that automatically detects and tracks horses and riders during equestrian events. A hybrid tracking approach maintains identification as horses move through the arena, while real-time tracking data controls camera movement to keep each horse in view. The system also automates video capture, creating individual horse clips without requiring a dedicated camera operator.

This mobile application streamlines inventory cataloging for teams working in the field. AI-assisted cataloging generates inventory information from images, while keeping users in control of reviewing and editing the results. The application also supports container-based inventory management, offline-first operation, auto-save, and synchronization helping field teams manage large volumes of items efficiently.

We developed a centralized portal to replace manual data entry, scattered tools, and email-based coordination with a connected digital workflow. AI-assisted document processing extracts information from tech packs and purchase orders, reducing manual data entry and connecting information directly to operational workflows. The platform brings style and order management, factory allocation, purchase contracts, document management, and logistics tracking together in one system.

We implemented a centralized AI gateway that allows applications to access multiple AI models through a single integration layer, without requiring changes to application code when models change. Configurable model routing and aliases make it easier to switch between models as requirements evolve. Virtual API keys, usage monitoring, budget limits, and model permissions provide controlled access across teams and applications, with visibility into AI usage and costs.

We developed a specialized platform that enables equestrian facilities to create and manage accurate digital barn layouts. Users can define stalls, aisles, dimensions, and other elements through an interactive designer. Multiple layouts can be saved, edited, and reused for different events. The platform also supports detailed visualization of stall locations, occupancy, and proximity to competition arenas. Administrators can reallocate groups, adjust timelines, and manage conflicts easily. AI capabilities for this platform use historical data and predictive analysis to improve stall allocation and adapt to changing event requirements.

We developed a custom authorization layer to secure enterprise MCP deployments and control access to AI-powered tools and data. The solution supports multi-tenant environments, role-based permissions, and user-specific authorization across different business applications. Layered authentication, request validation, rate limiting, permission controls, and comprehensive audit logging help protect each stage of the connection while providing visibility into access and tool activity.