I am an Assistant Professor in the Computer Science Department at University of Wisconsin, Madison. My research interests are in designing systems and algorithms for large scale data analysis and machine learning. My dissertation research looked at abstractions that make it easier to express new machine learning algorithms and systems that can improve their performance.
Before coming to Madison, I was a post-doctoral researcher in the Systems Research Group at Microsoft Research in Redmond. Previously, I completed my PhD from UC Berkeley where I was advised by Ion Stoica and Mike Franklin. I also have a Masters from University of Illinois at Urbana-Champaign and worked in the Systems Research Group, with Prof. Roy Campbell.
Teaching
CS 537 Intro to OS: F24 S23 S20 S19
CS 744 Big Data Systems: S25 S24 F22 F21 F20 F19 F18
CS 839: Advanced Machine Learning Systems: S22
Group
- Jason Mohoney (Phd Student, co-advised with Theodoros Rekatsinas)
- Konstantinos Kanellis (Phd Student)
- Rutwik Jain (Phd Student, co-advised with Matt Sinclair)
- Brandon Tran (Phd Student, co-advised with Matt Sinclair)
- Song Bian (Phd Student)
- Minghao Yan (Phd Student)
- Johannes Freischuetz (Phd Student)
- Tzu-Tao Chang (Phd Student)
- Fanchao Chen (Phd Student)
- Tareq Mahmood (Phd Student)
- Seth Ockerman (Phd Student)
Alumni
- Saurabh Agarwal (Phd to Post-doc at UT-Austin)
- Pengfei Zheng (Post-doc, co-advised with Aditya Akella)
- Aditi Singh (MS)
- Rachit Tibrewal (MS)
- Olesia Elfimova (MS, to Dropbox)
- Adarsh Kumar (MS, to Amazon Alexa AI)
- Arjun Balasubramanian (MS, to Amazon AWS)
- Wei Hao (BS, to Columbia)
- Yiheng Xu (BS, to Maryland)
- Yuhan Liu (BS, to UChicago)
- Ziyi Zhang (BS, to UChicago)
- Rui Pan (BS, to Princeton)
- Lynn Liu (BS, to UC Berkeley)
- Prasoon Sinha (BS, to UT Austin)
- Anze Xie (BS, to UCSD)
- Anders Carlsson (BS, to Amazon)
Recent Publications
Tzu-Tao Chang, Shivaram Venkataraman Eva: Cost-Efficient Cloud-Based Cluster Scheduling - Eurosys 2025
Johannes Freischuetz, Konstantinos Kanellis, Brian Kroth, Shivaram Venkataraman TUNA: Tuning Unstable and Noisy Cloud Applications - Eurosys 2025
Minghao Yan, Saurabh Agarwal, Shivaram Venkataraman Decoding Speculative Decoding - NAACL 2025
2024
Meguru Yamazaki, Shivaram Venkataraman CO2: Precise Attention Score Observation for improving KV Cache Replacement in Large Language Model - Efficient Systems for Foundation Models (ES-FoMO) Workshop at the International Conference on Machine Learning (ICML) 2024
Rutwik Jain, Brandon Tran, Keting Chen, Matthew Sinclair, Shivaram Venkataraman PAL: A Variability-Aware Policy for Scheduling ML Workloads in GPU Clusters - International Conference for High Performance Computing, Networking, Storage and Analysis (Supercomputing 2024)
Konstantinos Kanellis, Johannes Freischuetz, Shivaram Venkataraman Nautilus: A Benchmarking Platform for DBMS Knob Tuning - DEEM Workshop 2024
Saurabh Agarwal, Bilge Acun, Basil Homer, Mostafa Elhoushi, Yejin Lee, Shivaram Venkataraman, Dimitris Papailiopoulos, Carole-Jean Wu CHAI: Clustered Head Attention for Efficient LLM Inference - ICML 2024
Song Bian, Dacheng Li, Hongyi Wang, Eric Xing, Shivaram Venkataraman Does compressing activations help model parallel training? - MLSys 2024
Saurabh Agarwal, Amar Phanishayee, Shivaram Venkataraman Blox: A Modular Toolkit for Deep Learning Schedulers - Eurosys 2024
2023
Saurabh Agarwal, Chengpo Yan, Ziyi Zhang, Shivaram Venkataraman BagPipe: Accelerating Deep Recommendation Model Training - SOSP 2023
Qiyang Ding, Pengfei Zheng, Shreyas Kudari, Shivaram Venkataraman, Zhao Zhang Mirage: Towards Low-interruption Services on Batch GPU clusters with Reinforcement Learning - International Conference for High Performance Computing, Networking, Storage and Analysis (Supercomputing 2023)
Roger Waleffe, Jason Mohoney, Theodoros Rekatsinas, Shivaram Venkataraman MariusGNN: Resource-Efficient Out-of-Core Training of Graph Neural Networks - Eurosys 2023
Pengfei Zheng, Rui Pan, Tarannum Khan, Shivaram Venkataraman, Aditya Akella Shockwave: Fair and Efficient Cluster Scheduling for Dynamic Adaptation in Machine Learning - NSDI 2023
Harsh Darshan Sapra, Olesia Elfimova, Sahana Upadhya, Lukas Desorcy, Michael Wagner, Shivaram Venkataraman, Chol-Bum Kweon, Sage Kokjohn, Justin Shumaker Estimating Battery State-of-Charge within 1% using Machine Learning and Physics-based Models - SAE World Congress 2023
2022
Prasoon Sinha, Akhil Guliani, Rutwik Jain, Matthew Sinclair, Shivaram Venkataraman Not All GPUs Are Created Equal: Characterizing Variability in Large-Scale, Accelerator-Rich Systems - International Conference for High Performance Computing, Networking, Storage and Analysis (Supercomputing 2022)
Konstantinos Kanellis, Cong Ding, Brian Kroth, Andreas Müller, Carlo Curino, Shivaram Venkataraman LlamaTune: Sample-Efficient DBMS Configuration Tuning - VLDB 2022
Saurabh Agarwal, Hongyi Wang, Shivaram Venkataraman, Dimitris Papailiopoulos On the Utility of Gradient Compression in Distributed Training Systems - MLSys 2022
2021
Anze Xie, Anders Carlsson, Jason Mohoney, Roger Waleffe , Shanan Peters, Theodoros Rekatsinas, Shivaram Venkataraman Demonstration of Marius: Graph Embeddings with a Single Machine - VLDB 2021
Adarsh Kumar, Kausik Subramanian, Shivaram Venkataraman, Aditya Akella Doing more by doing less: how structured partial backpropagation improves deep learning clusters - DistributedML Workshop at CoNEXT 2021
Gregory Pauloski, Qi Huang, Lei Huang, Shivaram Venkataraman, Kyle Chard, Ian Foster, Zhao Zhang KAISA: An Adaptive Second-order Optimizer Framework for Deep Neural Networks - International Conference for High Performance Computing, Networking, Storage and Analysis (SC21)
Jason Mohoney, Roger Waleffe, Yiheng Xu, Theodoros Rekatsinas, Shivaram Venkataraman Marius: Learning Massive Graph Embeddings on a Single Machine - OSDI 2021
Arjun Singhvi, Arjun Balasubramanian, Kevin Houck, Mohammed Danish Shaikh, Shivaram Venkataraman, Aditya Akella Atoll: A Scalable Low-Latency Serverless Platform - SoCC 2021
Saurabh Agarwal, Hongyi Wang, Kangwook Lee, Shivaram Venkataraman, Dimitris Papailiopoulos Accordion: Adaptive Gradient Communication via Critical Learning Regime Identification - MLSys 2021
Le Xu, Shivaram Venkataraman, Indranil Gupta, Luo Mai and Rahul Potharaju Move Fast and Meet Deadlines: Fine-grained Real-time Stream Processing with Cameo - NSDI 2021
Yuhan Liu, Saurabh Agarwal, Shivaram Venkataraman AutoFreeze: Automatically Freezing Model Blocks to Accelerate Fine-tuning - arXiv preprint code
Arjun Balasubramanian, Adarsh Kumar, Yuhan Liu, Han Cao, Shivaram Venkataraman, Aditya Akella Accelerating Deep Learning Inference via Learned Caches - arXiv preprint
2020
Vaishaal Shankar, Karl Krauth, Kailas Vodrahalli, Qifan Pu, Ion Stoica, Benjamin Recht, Jonathan Ragan-Kelley, Eric Jonas, Shivaram Venkataraman Serverless Linear Algebra - SoCC 2020
Konstantinos Kanellis, Ramnatthan Alagappan, Shivaram Venkataraman. Too Many Knobs to Tune? Towards Faster Database Tuning by Pre-selecting Important Knobs - HotStorage 2020
Kshiteej Mahajan, Arjun Balasubramanian, Arjun Singhvi, Shivaram Venkataraman, and Aditya Akella, Amar Phanishayee, Shuchi Chawla. Themis: Fair and Efficient GPU Cluster Scheduling - NSDI 2020
Guanhua Wang, Shivaram Venkataraman, Amar Phanishayee, Nikhil Devanur, Jorgen Thelin, Ion Stoica Blink: Fast and Generic Collectives for Distributed ML - MLSys 20202019
Jack Kosaian, K.V. Rashmi, Shivaram Venkataraman Parity Models: Erasure-Coded Resilience for Prediction Serving Systems - SOSP 2019
Myeongjae Jeon, Shivaram Venkataraman, Amar Phanishayee, Junjie Qian, Wencong Xiao, Fan Yang Analysis of Large-Scale Multi-Tenant GPU Clusters for DNN Training Workloads - USENIX ATC 2019
John Emmons, Sadjad Fouladi, Ganesh Ananthanarayanan, Shivaram Venkataraman, Silvio Savarese, Keith Winstein Cracking open the DNN black-box: Video Analytics with DNNs across the Camera-Cloud Boundary - Hot Topics in Video Analytics and Intelligent Edges (HotEdgeVideo 2019)
Qifan Pu, Shivaram Venkataraman, Ion Stoica Shuffling, Fast and Slow: Scalable Analytics on Serverless Infrastructure - NSDI 2019
Please see here for a complete list.