Vikram Sharma Mailthody


Vikram Sharma Mailthody
Electrical and Computer Engineering Department
University of Illinois, Urbana Champaign

Office: Coordinate Science Laboratory, Room 227
Email: vsm2 at illinois dot edu
[LinkedIn] [Facebook] [Quora] [GitHub] [Google Scholar]

Short Bio

Vikram is a Research Scientist at NVIDIA Research. He classifies him as systems and platform researcher and a computer architect focusing on GPU memory/storage system design. He is interested in solving fundamental systems-level problems and proposing optimizations for emerging applications. Vikram has extensive experience in memory and storage system design, GPUs, and performance optimizations for emerging applications like NLP, recommender systems, GNNs, and graph and data analytics. Vikram holds a Ph.D. in Electrical and Computer Engineering from the University of Illinois, Urbana Champaign, in 2022. He completed his dissertation under the supervision of Professor Wen-mei Hwu at the IMPACT Research Group. In his dissertation, he discussed application support and adaptation for high-throughput accelerator orchestrated fine-grain storage access. Vikram is a recipient of the Bahl Fellowship for 2019–21 and the Dan Vivoli Endowed Fellowship for 2021–22 and has won several competitions, including the Championship award for the HPEC’20 Graph Challenge and multiple student innovation awards in competitions. His work has been featured in TheRegister, TomsHardware, PCGamer, ArsTechnica, and more.

Recent News

Aug 2022: Giving a tutorial talk on Sparisty in ML at MLSys’22. See you there!

Aug 2022: I have joined NVIDIA as Research Scientist.

Aug 2022: I graduated from UIUC!

Jun 2022: I defended my thesis. More about it will come in the future as the work is under submission!

Dec 2021: Two US patents granted!

Jul 2021: Received C3SR Best Research Award and C3SR Director's Distinguished Service Award for the work we did at IBM-C3SR initiative. Thank you everyone!

Apr 2021: Received Dan Vivoli Endowed Fellowship for the year 2021-22. Thank you!

Feb 2021: Organizing Special COVID-19 Session in CSLSC’21. [Link]

Jan 2021: Safer Illinois and RokWall: Privacy Preserving University Health Apps for COVID-19 is now available download.

Dec 2020: FlatFlash patent application is now published - US Patent No: US20200401530A1

Sept 2020: Our EMOGI work is accepted in Proceedings of the VLDB Endowment

Aug 2020: Our paper on secure-privacy preserving computing accepted in NDSS CoronaDef Workshop: Call for Innovative Secure IT Technologies against COVID-19. More details to come.

Aug 2020: Our winning source code of SparseDNN Challenge is now available here

Aug 2020: Won the Championship Award in Sparse DNN Graph Challenge, 2020! We proposed at-scale fast and efficient sparse DNN inference for GPU. News

Aug 2019: We got 2 awards in Graph Challenge, 2019 - Student Innovation Award for our work on k-truss decomposition and Honorable Mention Award for triangle counting!

July 2019: Our work DeepStore, supporting emerging deep-learning based data queries using in-storage acceleration was accepted in MICRO 2019!

Jun 2019: US Patent Accepted - US PubNo: US 2019/0197019 A1

Feb 2019: Received Bahl Fellowship (Photos) for the year 2019-21! Thank You!

Nov 2018: Our work FlatFlash, enabling large memory capacity using byte-accessible SSDs was accepted in ASPLOS 2019!

Aug 2018: We got 2 awards in Graph Challenge, 2018 - Graph Challenge Finalist and Student Innovation Award for our work on triangle counting in GPUs and FPGA, respectively!