Preyesh Dalmia

About Preyesh Dalmia

I\'m a Deep Learning architect at NVIDIA, where I work on optimizing LLM Pre-Training and Post-Training performance. Previously, I earned a PhD in Computer Engineering from the University of Wisconsin-Madison. My PhD research centered on data movement in heterogeneous systems. I have developed optimizations aimed at reducing synchronization and communication overheads in GPUs, from both software and hardware aspects. During my research, I examined these overheads across systems of varying scales, ranging from monolithic systems to those deployed in data centers.If your interests align with mine, or if you have an exciting opportunity, please feel free to connect!

Experience

1. Senior Deep Learning Architect

   NVIDIA

   Aug 2023 — Present · Santa Clara County, CA, US

   Enhancing Large Language Models (LLMs) Pre-Training and Post-Training performance by working on optimizations throughout the software and hardware stack

Education

• University of Wisconsin-Madison

  PhD, Computer Engineering

• Delhi College of Engineering

  Bachelor of Technology (B.Tech.), Electronics and Communication Engineering

  2012 — 2016

Skills

• Vhdl
• C++
• Creative Writing
• Fpga
• Asic
• Modelsim
• Corporate Communications
• Static Timing Analysis
• Public Speaking
• Digital Image Processing
• Logic Design
• Tcl-Tk
• Graphic Design
• Arduino
• Illustrator
• Presentation Skills
• C
• Verilog
• Pspice
• Xilinx Ise
• Microsoft Office
• Vlsi
• Very-Large-Scale Integration (Vlsi)
• Photoshop
• Technical Writing
• Matlab
• Indesign