Kevin Stephano

Software Engineering Manager for Deep Learning Frameworks at NVIDIA

San Francisco Bay Area

About

Experience

  • NVIDIA (Full-time · 13 yrs 7 mos)
    • Engineering Manager of PyTorch Compilers
      Jan 2026 - Present · 6 mos

    • Engineering Manager of Pytorch nvFuser Team
      Oct 2022 - Mar 2026 · 3 yrs 6 mos

    • Tech Lead / Software Engineer for Pytorch nvFuser Team
      Jun 2021 - Oct 2022 · 1 yr 5 mos

  • Advanced Micro Devices (5 yrs 8 mos)
    • Senior Design Engineer
      Apr 2010 - Dec 2012 · 2 yrs 9 mos

      Work in performance modeling and analysis of the core CPU.

    • Senior Design Engineer
      May 2007 - Mar 2010 · 2 yrs 11 mos

      Developed architectural model of the x86 - AMD64 instruction set architecture in the form of an execution driven simulator with a focus on the virtualization instructions. Additionally, worked for a short time on a random test generator.

  • Microprocessor Performance Intern at IBM
    May 2006 - Aug 2006 · 4 mos

    Ran the Spec 2006 on new architectural models looking for weaknesses. Investigated short branches that might be converted into predicated branches by the hardware without software support.

  • Microprocessor Performance intern at IBM
    May 2003 - Aug 2003 · 4 mos

    Worked on analyzing Fortran kernels for store-to-load forwarding and prefetching on the Power5 processor. Also, made compiler optimization recommendations.

  • FPU Logic Design Engineer at International Business Machines
    Sep 2000 - Aug 2002 · 2 yrs

    Floating point logic (RTL) designer for Power4, Power5, and PowerPC 970 microprocessors. Designed the power gating scheme for the Power5/PPC 970 floating point unit. Power designs were previously not power gated.