Jacob O. Torring

Hi! I work on the Inference Runtime at 1X Technologies, where I lead the end-to-end inference performance architecture for humanoid robots, owning everything that touches the GPU.

I did my PhD in High-Performance Computing (HPC) and AI at NTNU, with research focused on applying cutting-edge AI techniques to optimize HPC systems through autotuning.

I led the CATBench project with collaborators from Lund University and Stanford University, and I authored the BAT project, working with researchers from Leiden University and Masaryk University. These projects were community-driven efforts aimed at redefining autotuning evaluation in HPC. Additionally, I supervised the LS-CAT project, a large-scale CUDA autotuning dataset for training LLMs in CUDA kernel optimization. I supervised five Master’s students during my time at NTNU.

Previously, I worked as a Senior Software Engineer at Norway’s largest HPC center, where I developed a self-tuning analysis tool to characterize experimental systems using the Roofline Model. Prior to that, I was a Graduate Research Assistant at KAIST, where I developed a novel technique in probabilistic program slicing. I also led the DevOps team and co-developed the On-Board Computer of an operational satellite.

When I’m not immersed in the world of HPC and AI, you can find me playing Go at the Silicon Valley Go Club. I’m also a former freelance Lighting Designer for theater and concerts and founded a stage technician organization.

I’m passionate about pushing the boundaries of computational efficiency. Feel free to connect if you’re interested in the intersection of HPC and AI, or if you’d like to explore how my research can apply to your work!