Built at the intersection of supercomputing, GPU systems, parallel mathematics and infrastructure economics.

Dr. Kalyan Perumalla is a computer scientist who has guided multidisciplinary teams developing advanced computing technologies for government and national laboratories. He has directed a $100 million R&D portfolio for the U.S. government in AI, quantum computing and networking, supercomputing, and scientific discovery through advanced computing. He also developed algorithms and prototypes that have scaled to thousands of GPUs and 100s of thousands of processor cores. He has published scholarly articles on his research in international journals and conference proceedings. Dr. Perumalla earned his PhD in Computer Science from Georgia Tech. He has held senior faculty and adjunct positions at the University of Tennessee, Georgia Tech and the University of Nebraska, Lincoln and was an IAS Fellow at Durham University.

Dr. Perumalla brings his extensive high-end computing expertise to deliver on the Vectris mission to extract more value from AI infrastructures.

Vinod Tipparaju has held several senior engineering and architect roles involving high-performance computing (HPC) and graphics processors across AMD, Graphcore, Hewlett Packard Enterprise, and other places. He has also contributed to high-end computing for the U.S. as a senior computing scientist at national laboratories. He served as a Master Principal AI/GPU Architect at Oracle Cloud Infrastructure and as a GPU/AI Platform Architect at Bytedance. Vinod has published more than 70 scholarly papers and acquired 6 patents. He has led multiple R&D projects in topics covering HPC, GPU, ML, and compilers, resulting in significant performance improvements and funding opportunities.

Vinod builds on his extensive expertise in hardware and software co-design, parallel systems, performance measurement and runtime environments in creating products and systems that help Vectris lead the way in developing the field of AI compute yield economics.

Dr. Joel Malard has more than three decades of experience building high integrity systems for large-scale computing and optimization. Dr. Malard has deep expertise in computational mathematics and software spanning sparse linear algebra, adjoint methods, multi-objective optimization (Pareto ranking), parallel and distributed computing, simulation of complex systems including power networks, market dynamics (Black-Scholes-Merton), and circuit level modeling. He holds a PhD from McGill University.

Dr. Malard leads the engineering of bounded runtime and certification layers that enable infrastructure teams to recover high-value capacity from existing AI infrastructures. The process turns wasted compute into measurable gains in usable capacity, margins and revenue without retraining models or changing hardware.