Scholar
Dan Alistarh
Google Scholar ID: 75q-6ZQAAAAJ
Professor at IST Austria
Citations & Impact
Citations
12,822
H-index
40
i10-index
99
Publications
20
Co-authors
24
list available
Publications
Resume (English only)
Academic Achievements
- Supported by the Austrian FWF Center of Excellence BILAI
- Recipient of ERC Proof-of-Concept and Starting Grants
- Received generous grants from NVIDIA, Google, and Amazon
- 5 papers accepted to NeurIPS 2025, including one Spotlight: HALO, Compression Scaling Laws, Hogwild! Inference, Quartet, Influence Distillation
- 4 papers accepted to ICML 2025: AQUA-KV, EvoPress, QuEST, Optimistic Dual Averaging
- 4 papers accepted to NeurIPS 2024, including one Oral: PV-Tuning, microAdam, Iterative OBS, QuaRot
- Paper on Marlin (quantized kernels for mixed-precision LLM inference) accepted to PPoPP 2025
- Two papers accepted to EMNLP 2024: MathadorLM (benchmarking LLM math skills) and QUIK (weight and activation quantization for LLMs)
- 4 papers accepted to ICML 2024: SPADE, AQLM, RoSA, compressed preconditioners
- Two papers accepted to MLSys 2024: QMoE, L-GreCo
- Two papers presented at ICLR 2024: Scaling laws for sparse models (Spotlight), SpQR
- Three papers at NeurIPS 2023: CAP (unstructured pruning), Variance-Reduction interpretation of Knowledge Distillation, ZipLM (structured pruning)
- Three papers accepted to ICML 2023: SparseGPT, SparseProp, QSDP
- Supervised Elias Frantar’s influential work on LLM compression including GPTQ, Marlin, SparseGPT, QMoE; GPTQ models downloaded millions of times on HuggingFace
Research Experience
- Professor at the Institute of Science and Technology Austria (ISTA)
- ML Research Lead at Neural Magic, Inc.
- Postdoctoral Associate at MIT CSAIL, working with Prof. Nir Shavit
- Researcher at Microsoft Research, Cambridge, UK
- Researcher at ETH Zurich
- Visiting Professor at MIT during Fall 2023
