Suggestions
Karu Sankaralingam
Principal Research Scientist at NVIDIA and Professor at UW Madison
Karu Sankaralingam is a Principal Research Scientist at NVIDIA and a Professor at the University of Wisconsin-Madison.12 He has been working at NVIDIA since August 2021 and has held a professorship at UW-Madison since April 2017.1
Sankaralingam's academic background includes:
- B.Tech in Aerospace Engineering from the Indian Institute of Technology, Madras (1995-1999)
- PhD in Computer Science from the University of Texas at Austin (1999-2006)1
His career highlights include:
- Co-founder, CEO & CTO of SimpleMachines, Inc. (January 2017 - August 2021)
- Visiting Professor at Qualcomm (July 2014 - December 2014)
- Associate Professor at UW-Madison (January 2007 - Present)1
Sankaralingam is known for his research in advanced microprocessor technologies.1 He has led three significant chip projects: Mozart (16nm, HBM2 based design), MIAOW open source GPU on FPGA, and the TRIPS chip during his PhD.3 His work has been featured in industry forums and covered by major publications like the New York Times, Wired, and IEEE Spectrum.3
Sankaralingam has received several awards, including:
- IEEE Fellow
- Vilas Faculty Early Career Investigator Award (2018)
- Wisconsin Innovation Award (2016)
- IEEE TCCA Young Computer Architecture Award (2012)
- NSF CAREER award (2009)3
His research focuses on dataflow computing, architecture, microarchitecture, and compiler principles.3 Sankaralingam has published over 100 research papers, graduated 9 PhD students, and is an inventor on 21 patents.3
Highlights
Having spent years conducting PC reviews and engaging in online discussions, I've come to a perplexing realization: my frustration often stems not from positive or negative feedback on my papers, but from the sheer act of reading others' reviews. Interestingly, in SIGARCH conferences, our current triple-blind review process—where authors, reviewers, and peer reviewers remain anonymous until PC discussions—creates a peculiar dynamic. Approximately 40% of submissions reach these discussions, allowing a brief glimpse into the identities behind the reviews. For the remaining 60% this anonymity paradoxically fosters a more objective and less emotionally charged evaluation of the content. Without any knowledge of the names it’s easy to objectively critique the review and label for the behaviors that are happening either implicitly or explicitly in reviewers’ minds.
My humorous take on this. At some level I am stating the obvious and this isn't novel. Thoughts?
Thrilled to announce public availability of SimpleMachines' first chip Mozart : TSMC 16nm chip with HBM2 memories. It's optimized for inference and can run Tensorflow out-of-the-box. And we can run up to 64 different AI models simultaneously on one chip: we've condensed an entire rack of compute to one chip. We are closely working with our lead customers gearing up for production. 5 MLPerf models available for immediate customer use. Contact us for trials and running your own models. Kudos to our amazing team. #ai #semiconductors #deeplearningai #hardware
https://lnkd.in/eAXKgSv
