Casper Wilstrup

Spectral brings the paradigm of CPU programming over to HPC computing platforms: write once, recompile for different hardware. We develop SCALE, a toolkit comprising of a compiler, CUDA Language Extensions and open source libraries providing the "CUDA-X" APIs (e.g. cuBLAS, cuSOLVER, etc). It compiles a superset of CUDA to AMD machine code, with no intrinsic overhead, allowing engineers and data scientists to choose the hardware platform best suited to their HPC/AI/ML needs.

Yaak accelerates the adoption of end-to-end (E2E) learning for spatial intelligence by building essential infrastructure: tools for dataset visualization and search, the world’s largest open-source multimodal dataset, and libraries for training and deploying E2E AI across robotics verticals. A natural language interface streamlines the entire workflow—allowing both experts and non-technical users to train and deploy robotics AI with ease. This enables a new paradigm: using English as the interface for programming spatially intelligent robots.

Stop making disconnected applets, and start building yourself a connected toolbox.

Spectral brings the paradigm of CPU programming over to HPC computing platforms: write once, recompile for different hardware. We develop SCALE, a toolkit comprising of a compiler, CUDA Language Extensions and open source libraries providing the "CUDA-X" APIs (e.g. cuBLAS, cuSOLVER, etc). It compiles a superset of CUDA to AMD machine code, with no intrinsic overhead, allowing engineers and data scientists to choose the hardware platform best suited to their HPC/AI/ML needs.

Yaak accelerates the adoption of end-to-end (E2E) learning for spatial intelligence by building essential infrastructure: tools for dataset visualization and search, the world’s largest open-source multimodal dataset, and libraries for training and deploying E2E AI across robotics verticals. A natural language interface streamlines the entire workflow—allowing both experts and non-technical users to train and deploy robotics AI with ease. This enables a new paradigm: using English as the interface for programming spatially intelligent robots.

Aves helps companies understand AI and carve out their place in the technology landscape of tomorrow.

Abzu is the best in class in silico disease understanding and drug design company. Since its founding in 2018, Abzu’s proprietary explainable AI, the QLattice®, has been the cornerstone to their cutting-edge approach. Abzu is a preferred partner for pharma, biotech, and research institutions, transforming data into insights and accelerating R&D. Abzu has offices in Copenhagen and Barcelona, and has raised €13.3M in total funding to date. Abzu has been mentioned in Gartner® research, from Market Guides for AI Trust, Risk, and Security Management to Cool Vendor™, and is the winner of several awards, including the 2023 “Best HealthTech” Nordics Award at the Nordic Startup Awards.