Shivaram Venkataraman
Associate Professor
Builds systems for machine learning and large-scale data analytics, including scheduling and resource allocation for distributed model training in the cloud.
Shivaram Venkataraman's lab at ETH Zurich works on distributed, cloud & edge systems, efficient training & inference and data management & big data. 28 papers since 2023, 15 at top venues, mostly EuroSys and ICML. 12 alumni, most often to Amazon and NVIDIA.
At a glance
- papers, 2023–2026
- 7/ year
- at T1 venues
- 54%
- PhD students
- 0
- MS students
- 0
- alumni listed
- 12
Papers by yearfrom DBLP, 2023–2026
- 8 papers2026 · T1 2 · collecting
- 11 papers2025 · T1 7
- 5 papers2024 · T1 3
- 4 papers2023 · T1 3
T1 venuesOtherVenues: EuroSys, ICML, NSDI, ICLR, MLSys
Research focuspaper titles that use each topic's words
- Distributed, cloud & edge systems10 papers
- Efficient training & inference9 papers
- Data management & big data1 paper
- allocation0 papers
- resource allocation0 papers
Paper list
From DBLP.
20268 papers
- SIGMOD CompanionDemo of SemWeave: Semantic Common Expressions for LLM-powered Query Processing (opens in a new tab)
- IPDPSWET: Bridging the Gap on Energy Telemetry for Multi-GPU Communication Collectives (opens in a new tab)
- IEEE Trans. ComputersFrom Good to Great: Parameter Tuning in Memory Tiering Systems (opens in a new tab)
- SIGMETRICSMinos: Systematically Classifying Performance and Power Characteristics of GPU Workloads on HPC Clusters (opens in a new tab)
- DIMES@SOSPRegional Policies for Memory Tiering Systems (opens in a new tab)
- ICLR (T1)Scaling Laws Meet Model Architecture: Toward Inference-Efficient LLMs (opens in a new tab)
- NSDI (T1)SYMPHONY: Enabling Compute-Memory Disaggregation in LLM Serving Systems (opens in a new tab)
- ICSWattchmen: Watching the Wattchers - High Fidelity, Flexible GPU Energy Modeling (opens in a new tab)
202511 papers
- NAACL (T1)Decoding Speculative Decoding (opens in a new tab)
- eScienceDiamond: Harnessing GPU Resources for Scientific Deep Learning (opens in a new tab)
- EuroSys (T1)Eva: Cost-Efficient Cloud-Based Cluster Scheduling (opens in a new tab)
- SC WorkshopsExploring Distributed Vector Databases Performance on HPC Platforms: A Study with Qdrant (opens in a new tab)
- VLDB (T1)From FASTER to F2: Evolving Concurrent Key-Value Store Designs for Large Skewed Workloads (opens in a new tab)
- ICML (T1)LV-XAttn: Distributed Cross-Attention for Long Visual Inputs in Multimodal Large Language Models (opens in a new tab)
- SCPGT-I: Scaling Spatiotemporal GNNs with Memory-Efficient Distributed Training (opens in a new tab)
- OSDI (T1)Quake: Adaptive Indexing for Vector Search (opens in a new tab)
- ICML (T1)Scaling Inference-Efficient Language Models (opens in a new tab)
- DIMES@SOSPStriking the Right Chord: Parameter Tuning in Memory Tiering Systems (opens in a new tab)
- EuroSys (T1)TUNA: Tuning Unstable and Noisy Cloud Applications (opens in a new tab)
20245 papers
- EuroSys (T1)Blox: A Modular Toolkit for Deep Learning Schedulers (opens in a new tab)
- ICML (T1)CHAI: Clustered Head Attention for Efficient LLM Inference (opens in a new tab)
- MLSys (T1)Does Compressing Activations Help Model Parallel Training? (opens in a new tab)
- DEEM@SIGMODNautilus: A Benchmarking Platform for DBMS Knob Tuning (opens in a new tab)
- SCPAL: A Variability-Aware Policy for Scheduling ML Workloads in GPU Clusters (opens in a new tab)
20234 papers
- SOSP (T1)Bagpipe: Accelerating Deep Recommendation Model Training (opens in a new tab)
- EuroSys (T1)MariusGNN: Resource-Efficient Out-of-Core Training of Graph Neural Networks (opens in a new tab)
- SCMirage: Towards Low-interruption Services on Batch GPU Clusters with Reinforcement Learning (opens in a new tab)
- NSDI (T1)Shockwave: Fair and Efficient Cluster Scheduling for Dynamic Adaptation in Machine Learning (opens in a new tab)
Where alumni went12 listed, 71% to industry
Industry 5Academia 2
Most often to
- Amazon
- NVIDIA
Also to
- MIT
Apply to ETH ZurichRolling
Not announced yet
Computer Science Doctorate (D-INFK)
- Round
- Rolling (no deadline)
- Opens
- —
- GRE
- See program page
- English
- Proficiency in written and spoken English
- Fee
- See program page
Confirmed on the program page.
Similar labsnearest by research summary
- CNZhejiang
Xiaohang Wang
- Accelerators, chips & architecture
- SGNUS
Bingsheng He
- Accelerators, chips & architecture
- USUCSD
Yiying Zhang
- Efficient training & inference
- USMIT
Tim Kraska
- Large language models
- CNPeking
Xin Jin
- Efficient training & inference
- CNSJTU
Zeyu Mi
- Efficient training & inference
Your notes
Save this lab to keep notes.
Data
Checked against the lab homepage.
- Collected
- Verified
- Papers
- DBLP
- Members
- Lab homepage
- Alumni
- Lab homepage
Tiers are venue classifications, not lab rankings.