dividiti / cTuning foundation
Interdisciplinary researchers often struggle to apply novel techniques developed for one domain to another due to a lack of common tools, interfaces and meta-information. Worse, they waste increasing amounts of time trying to adapt to continuously evolving software and hardware, and rapidly growing amounts of data.
We provide highly effective co-design and optimization of innovative products across the full AI/SW/HW stack in terms of of speed, accuracy, energy, size, complexity, costs and other metrics. We pioneered a unique workflow technology to solve above problems and automatically assemble (co-design) the most efficient algorithm/software/hardware stacks from continuously optimized Collective Knowledge components (models, dataset,s algorithms, frameworks, libraries, compilers) for emerging customer workloads (AI,ML,quantum) in terms of speed, accuracy, energy, resilience and costs from cloud to edge.
Our clients include leading global semiconductor and automotive companies.