Fort Worth, Texas – May 2025
Graiphic had the honor of presenting its SOTA framework during the closing keynote of Norman Kirchner Jr. at NI Connect 2025. This was a major milestone for our team, highlighting years of intense development, driven by a clear mission: deliver a complete and unified deep learning ecosystem embedded directly into LabVIEW.
Our presentation was very well received by the audience, confirming that our vision resonates with both industrial and academic communities. We will continue our efforts to raise awareness of Graiphic’s technologies in the North American market.
Introducing SOTA: Much More Than a Toolkit
SOTA is not just a deep learning toolkit. It is a fully integrated framework designed to simplify the development and deployment of AI applications without compromising performance or scalability.
During our talk, we introduced the entire architecture behind SOTA, which includes:
-
Specialized toolkits for deep learning, computer vision, generative AI, GPU acceleration and annotation
-
Tools for model import, license management, one-click installation, and environment control
-
A single execution backend powered by ONNX Runtime, compatible with CPU, GPU, FPGA and edge platforms
Why LabVIEW and ONNX Runtime?
Our approach is based on a strong technological foundation. We chose LabVIEW for its graphical abstraction capabilities and its proven industrial maturity. Combined with ONNX Runtime – an efficient, cross-platform inference engine – we deliver a seamless experience from model development to deployment.
Graiphic is a proud contributor to ONNX and the only European company actively involved in the evolution of this open source standard. This position strengthens our capacity to innovate on both the execution and interoperability fronts.
An Alternative to Fragmented Python Ecosystems
The current AI landscape is saturated with disconnected Python libraries, tools, and dependencies. Reproducibility, stability and maintainability are often compromised. In contrast, SOTA offers a tightly integrated and cohesive environment, eliminating the need for external scripting or containerization.
This unified approach streamlines workflows, reduces overhead, and makes AI accessible to engineers, researchers, and system integrators who need reliability and performance.
A Response to the Complexity Highlighted in MAD 2025
In light of the MAD 2025 (Machine Learning, AI and Data Landscape) report, it is clear that the fragmentation of AI tools is becoming a major challenge for professionals. Our vision with SOTA is to address this head-on by delivering a modular, consistent, and scalable framework that bridges the gap between cutting-edge AI and real-world engineering.
We are not building yet another Python wrapper. We are building a native, visual, and extensible ecosystem for the next generation of AI-powered systems.
