Luma seeks a seasoned performance engineer to accelerate multimodal models across GPU, CPU, and accelerators. You will write high‑performing PyTorch, Triton, and CUDA kernels and push hardware to the limit while preserving model quality.
You’ll own profiling, optimization, and deployment at scale, develop fused kernels, tensor-core aware code, and build monitoring tools for distributed training and inference.
#J-18808-Ljbffr
GPU Kernel & Transformer Performance Scientist employer: Lumaai
Luma is an exceptional employer for Creative Technologists, offering a dynamic work environment where innovation meets hands-on creativity. With a strong focus on employee growth and collaboration, team members are encouraged to push the boundaries of generative AI while working closely with sales and creative teams. Located in the UK, Luma fosters a culture of inclusivity and support, ensuring that every employee has the opportunity to thrive and make a meaningful impact in the world of technology and creativity.
