Author of "KernelEvolve: Scaling Agentic Kernel Coding for Heterogeneous AI Accelerators at" on arXiv
How media typically covers Michael Kuchnik
Michael Kuchnik as author
KernelEvolve, an agentic framework for automated kernel generation and optimization, reduces recommendation model optimization development time from weeks to hours while achieving 100% correctness across 250 benchmark problems on heterogeneous AI accelerators.
“Author of "KernelEvolve: Scaling Agentic Kernel Coding for Heterogeneous AI Accelerators at" on arXiv”