Santosh Pandey

PhD Candidate, Rutgers University

headshot.jpeg

Room 535, CoRE Building

Piscataway, NJ

I am a PhD candidate in Computer Engineering at Rutgers University, supervised by Prof. Hang Liu. I completed my undergraduate from Tribhuvan University, Nepal where my undergraduate research advisor was Prof. Subarna Shakya. During my PhD, I interned at Google, Lawerence Berkeley National Lab and Brookhaven National Lab.

My research interests lie at the intersection of systems and ML/HPC applications, focusing on full-circle innovation–using ML/HPC to design better systems (more current focus) and building optimized systems for ML/HPC applications. Notably, I have worked on developing accurate and reusable ML-based microarchitecture simulators while also accelerating and scaling these simulation models with hardware-software co-design, demonstrating the synergy between system optimization and ML/HPC.

My long-term focus is on automating the design and optimization of computing stacks (programs, compilers, and microarchitecture), leveraging generative models for design generation and deep learning models for performance modeling/optimization.


I enjoy hiking and playing soccer in my free time.


ExchangeMeet: Diverse social between two labs from any field, something I started in our lab (learnings/blogs).


[ I will be graduating in May 2025, and I am in the job market for tenure-track faculty positions in academia. ]

Resume    CV

News

Mar 21, 2025 One paper Concorde: Fast and Accurate CPU Performance Modeling with Compositional Analytical-ML Fusion, led by Arash, accepted at ISCA’25. A great collaborative work during my time at Google. party-popper
May 15, 2024 I will be joining Google for the summer as a student researcher working on ML methods to accelerate architecture simulation.   google-logo
Apr 11, 2024 TAO: Rethinking DL-based microarchitecture simulation is accepted at SIGMETRIC’24 journal (Acceptance rate ~11%). A great collaboration with Amir from Google DeepMind. party-popper
Mar 07, 2024 TEA+, led by Dr. Chengying Huan, accepted at TACO’24 journal.   party-popper
Sep 15, 2022 One paper accepted at SC’22 conference on accelerating ML-based microarchitecture simulation.   party-popper

Selected publications

  1. ISCA
    Concorde: Fast and Accurate CPU Performance Modeling with Compositional Analytical-ML Fusion
    In International Symposium on Computer Architecture (ISCA), 2025
  2. SIGMETRICS
    TAO: Re-Thinking DL-based Microarchitecture Simulation
    Santosh Pandey, Amir Yazdanbakhsh, and Hang Liu
    Proceedings of the ACM on Measurement and Analysis of Computing Systems , 2024
  3. SC
    C-SAW: A framework for graph sampling and random walk on GPUs
    Santosh Pandey, Lingda Li, Adolfy Hoisie, and 2 more authors
    In SC20: International Conference for High Performance Computing, Networking, Storage and Analysis , 2020
  4. TPDS
    Trust: Triangle counting reloaded on GPUs
    Santosh Pandey, Zhibin Wang, Sheng Zhong, and 8 more authors
    IEEE Transactions on Parallel and Distributed Systems , 2021
  5. SC
    Scalable deep learning-based microarchitecture simulation on GPUs
    Santosh Pandey, Lingda Li, Thomas Flynn, and 2 more authors
    In SC22: International Conference for High Performance Computing, Networking, Storage and Analysis , 2022
  6. HPEC
    H-index: Hash-indexing for parallel triangle counting on GPUs
    Santosh Pandey, Xiaoye Sherry Li, Aydin Buluc, and 2 more authors
    In 2019 IEEE high performance extreme computing conference (HPEC) , 2019