🧑‍🔬 Tri Dao

🏛️ Affiliations

🎓Princeton UniversityPrinceton · United States2023–2026 · 23 paperscurrent🤝TogetherLacey · United States2025–2026 · 2 paperscurrent🎓Stanford UniversityStanford · United States2017–2023 · 27 papers🎓University at Buffalo, State University of New YorkBuffalo · United States2018–2021 · 4 papers

From the bylines of their papers (OpenAlex, 2026-10-14), ORCID and Wikidata; institutions on papers are matched by machine. 2 more institutions on single bylines are not listed.

🧭 Fields, topics and concepts

Of its works in our library, with the number of works on each.

📅 Works by year

2020s

📄 Papers

The most cited first (OpenAlex and our library).

1. FlashAttention-2: Faster Attention with Better Parallelism and Work Partitioning (2023 · 1,240 citations)
   The time and memory required by the attention layer in a Transformer grow quadratically with sequence length.
2. FlashAttention-3: Fast and Accurate Attention with Asynchrony and Low-precision (2024 · 400 citations)
   The attention mechanism is the main computational bottleneck of the Transformer architecture: its cost grows quadratically with sequence length.
3. Medusa: Simple LLM Inference Acceleration Framework with Multiple Decoding Heads (2024 · 278 citations)
   Generating text with an LLM is slow because tokens come out one at a time. Each step has to read all of the model's weights from HBM, so the accelerator spends most of its time waiting on memory rather than computing. Speculative decoding helps, but it needs a separate draft model, which is costly to obtain and…

All works

20 works of theirs in our library, the most cited first; 3 with a page here, the others link their original until we summarise them.

1. FlashAttention: Fast and Memory-Efficient Exact Attention with IO-Awareness ↗arXiv:2205.14135 · no summary yetTri Dao et al.2022 · 2,495 citations
2. Mamba: Linear-Time Sequence Modeling with Selective State Spaces ↗arXiv:2312.00752 · no summary yetAlbert Gu, Tri Dao2023 · 1,726 citations
3. FlashAttention-2: Faster Attention with Better Parallelism and Work PartitioningarXiv:2307.08691Tri Dao2023 · 1,240 citations
4. StarCoder: may the source be with you! ↗arXiv:2305.06161 · no summary yetRaymond Li et al.2023 · 732 citations
5. FlashAttention-3: Fast and Accurate Attention with Asynchrony and Low-precisionConference on Neural Information Processing Systems (NeurIPS)Jay Shah, Ganesh Bikshandi, Ying Zhang, et al.2024 · 400 citations
6. Medusa: Simple LLM Inference Acceleration Framework with Multiple Decoding HeadsarXiv:2401.10774Tianle Cai, Yuhong Li, Zhengyang Geng, et al.2024 · 278 citations
7. HiPPO: Recurrent Memory with Optimal Polynomial Projections ↗arXiv:2008.07669 · no summary yetAlbert Gu et al.2020 · 205 citations
8. Deja Vu: Contextual Sparsity for Efficient LLMs at Inference Time ↗ICML 2023 · no summary yetZichang Liu et al.2023 · 185 citations
9. Hyena Hierarchy: Towards Larger Convolutional Language Models ↗arXiv:2302.10866 · no summary yetMichael Poli et al.2023 · 182 citations
10. An Empirical Study of Mamba-based Language Models ↗arXiv:2406.07887 · no summary yetRoger Waleffe et al.2024 · 67 citations
11. The Mamba in the Llama: Distilling and Accelerating Hybrid Models ↗arXiv:2408.15237 · no summary yetJunxiong Wang; Daniele Paliotta; Avner May; Alexander M. Rush;, Tri Dao2024 · 59 citations
12. Decentralized Training of Foundation Models in Heterogeneous Environments ↗arXiv:2206.01288 · no summary yetBinhang Yuan et al.2022 · 59 citations
13. SonicMoE: Accelerating MoE with IO and Tile-aware Optimizations ↗arXiv:2512.14080 · no summary yetWentao Guo et al.2025 · 14 citations
14. Adaptive Hashing for Model Counting. ↗Uncertainty in Artificial Intelligence · no summary yetJonathan Kuck et al.2019 · 7 citations
15. Beat the long tail: Distribution-Aware Speculative Decoding for RL Training ↗arXiv (Cornell University) · no summary yetShao, Zelei et al.2025 · 6 citations
16. Transform Once: Efficient Operator Learning in Frequency Domain ↗neural information processing systems · no summary yetMichael Poli et al.2022 · 5 citations
17. Approximating the Permanent by Sampling from Adaptive Partitions ↗arXiv (Cornell University) · no summary yetJonathan Kuck et al.2019 · 5 citations
18. ButterflyFlow: Building Invertible Layers with Butterfly Matrices ↗arXiv (Cornell University) · no summary yetChenlin Meng et al.2022 · 2 citations
19. AI+HW 2035: Shaping the Next Decade ↗arXiv (Cornell University) · no summary yetDeming Chen et al.2026 · 1 citation
20. When RL Meets Adaptive Speculative Training: A Unified Training-Serving System ↗arXiv (Cornell University) · no summary yetJunxiong Wang et al.2026

🤝 Co-authors here

4 co-authors with a page, the most papers together first.

• Deming Chen
• Tianle Cai
• Ying Zhang
• Yuhong Li

Venues

Where their works here were published, the most works first, with the editions of those that have a page.

• Conference on Neural Information Processing Systems (NeurIPS) ×2: NeurIPS 2024
• ICML 2023
• Uncertainty in Artificial Intelligence
• neural information processing systems

🔗 Cited by

Papers of our library cite their works 5,065 times (of the 149,439 papers whose bibliographies we read); 462 citing papers with a page here →.

Updated 30 October 2026

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/en/authors/tri-dao/