Industrial edge AI specialist Numurus has joined the NVIDIA Metropolis programme to help bring new industrial vision AI and robotics applications to market. The programme has an ecosystem and developer tools designed for industrial vision AI.

According to Numurus, researchers and application solution providers spend up to 40% or more of their total software budget building and maintaining low-level software, drivers, and tools. This effort is non-core for most development teams but a required component of any smart system software stack. Providers face major challenges building low-level software with hard-to-find expertise, short timelines, small budgets, and the high effort of integrating and maintaining sensor drivers, control system interfaces, communications drivers, tools for managing AI at the edge, automation, and beyond. “Providers need to avoid this complex, costly approach and use a scalable path to meet and grow demand for industrial automation,” said Shanif Merchant, Numurus chief product officer.

Numurus aims to address these challenges with its NEPI OS platform combined with off-the-shelf edge-compute hardware and engineering services. NEPI is a commercial, open-source, turnkey OS with plug-and-play abstracted sensor and communication drivers, backseat robotic and manipulator interfaces, drag-and-drop AI and automation tools, and local and remote system and data management solutions that reduce costs and simplify smart system development.

Numurus incorporates NVIDIA CUDA-X software libraries, the JetPack development environment, Nsight developer tools, and Jetson system-on-modules within its product line. Future integration with the NVIDIA TAO Toolkit and DeepStream SDK will enable developers to build their AI models and manage datasets even more rapidly.