Tri Dao

Co-Founder & Chief Scientist, Together AI; Assistant Professor of Computer Science, Princeton University at together-ai

Reviewed Updated Jul 28, 2026

This profile is AI-generated. If you spot an error, please help us fix it by sharing a URL to the correct information.

Location Princeton, NJ
Last Verified Investment Sail Research (Series A) — Jun 24, 2026
Social @tri_dao
Stage Focus
seed series-a
Sector Focus
AI machine-learning ai-infrastructure developer-tools

Background

Tri Dao is a co-founder and Chief Scientist of Together AI, and an Assistant Professor of Computer Science at Princeton University, where he leads the Dao AI Lab 12. He completed his PhD in Computer Science at Stanford, co-advised by Christopher Ré and Stefano Ermon, before joining Princeton in 2024 31.

Dao is widely known in the ML systems community for FlashAttention (2022) and Mamba (2023), both of which have become standard components in modern LLM training and inference stacks 24. His research focuses on hardware-aware algorithms and sequence models with long-range memory 12. He was named an AI2050 Early Career Fellow by Schmidt Sciences in 2025 4.

Together AI was founded in June 2022 by Vipul Ved Prakash (CEO), Ce Zhang (CTO), Chris Ré, Percy Liang, and Tri Dao 5. Dao joined as Chief Scientist in July 2023, with a stated goal of “making open source AI more accessible and cost-competitive” 3. The company reached an $8.3 billion valuation following an $800M Series C led by Aramco Ventures in 2026 5.

Dao’s angel investing activity is a small, personal side of a research- and operator-first career. He is not a full-time investor and does not run a fund. Publicly verifiable checks are concentrated in AI infrastructure and applied ML tooling — companies close to his technical expertise.

Stated Thesis

Dao has not published a written investment thesis. His public writing and interviews focus on ML research (efficient training and inference, hardware-aware algorithms, sequence modeling) rather than investing philosophy 16. Where he has spoken about the AI stack more broadly, he has emphasized the importance of open-source infrastructure and efficient inference:

  • On joining Together AI: “I’m excited to announce that I’m joining Together AI as Chief Scientist, with the goal of making open source AI more accessible and cost-competitive” 3.
  • On his research direction: “We plan to push on research to make training, fine-tuning, and inference as efficient and widely available as possible” 3.

Beyond that, no self-reported thesis for angel investments has been located in public sources.

Inferred Thesis

Sample size caveat. Only a small number of Dao’s angel investments have been independently confirmed via press releases or company announcements (4 rounds as of July 2026). Percentages are not meaningful at this sample size — the description below is qualitative.

Across confirmed checks, the pattern is consistent:

  • Sector: 100% AI / ML infrastructure and applied ML tooling. All four verified rounds are companies whose core product is training, fine-tuning, or inference infrastructure for LLMs and AI agents 78910.
  • Stage: Seed and Series A. Dao appears as one of several named individual angels — not as a lead — alongside institutional VCs and other technical operator-angels 78910.
  • Founder profile: Technically credentialed founders, several with prior research or infrastructure engineering pedigree (Adaptive ML was founded by the Falcon LLM team; Prime Intellect is a decentralized training platform; Sail Research is inference infrastructure for long-horizon agents) 7810.
  • Co-investor pattern: Frequently appears in rounds that also include other technical operator-angels — Dylan Patel (SemiAnalysis), Andrej Karpathy, Clem Delangue (Hugging Face), Thomas Wolf (Hugging Face), Elad Hazan (Princeton/Google DeepMind), John Hennessy (Alphabet), and Lip-Bu Tan (Intel) 78910. This suggests his angel dealflow overlaps with the ML researcher / infrastructure operator community.
  • Geography: Primarily US-based companies; Adaptive ML is Paris/NY-based 8.

Notable gap: No verified checks into applied consumer AI, vertical SaaS, or non-AI sectors. Dao’s confirmed public angel activity is tightly clustered around technical infrastructure he can directly evaluate.

Check size is not disclosed in any of the announcements reviewed.

Portfolio

Only investments where Tri Dao is explicitly named as an individual angel in the primary announcement are included. Corporate investments made by Together AI are excluded.

Company Year Stage Source
Adaptive ML 2024 Seed ($20M, led by Index Ventures) 7
Prime Intellect 2025 Seed extension ($15M, led by Founders Fund) 9
Dedalus Labs 2025 Seed ($11M, co-led by Kindred Ventures and Saga Ventures) 10
Sail Research 2026 Series A ($80M combined seed + Series A; Sequoia seed, Kleiner Perkins Series A) 811

Only 4 of the 5 investments listed on Dao’s Crunchbase profile have been independently verified via primary sources 12. The fifth is not linked in public press releases reviewed.

In Their Own Words

The following quotes are from Dao’s public statements about his research and the AI stack — not about investing specifically, which he does not publicly discuss.

On the durability of the Transformer:

“Transformer is still incredibly strong, very well supported, both hardware and software” — Tri Dao, Interconnects podcast interview with Nathan Lambert, December 21, 2023 6.

On state-space models and Mamba:

“We wanted to show that state space can be competitive or maybe even meet some of the transformers out there.” — Tri Dao, Interconnects, December 21, 2023 6.

On his research approach:

“I’ve been working at the intersection of machine learning and systems, so designing algorithms that take advantage of the hardware.” — Tri Dao, Interconnects, December 21, 2023 6.

On what matters most for model performance:

“Data is still the most important thing.” — Tri Dao, Interconnects, December 21, 2023 6.

On his role at Together AI when he joined:

“I’m excited to announce that I’m joining Together AI as Chief Scientist, with the goal of making open source AI more accessible and cost-competitive.” — Tri Dao, Together AI blog, July 17, 2023 3.

What Founders Say

No independently sourced founder testimonials specifically about Tri Dao’s work as an angel investor have been located. Portfolio company announcements name him among their angels but do not include founder statements about his post-investment involvement.

Sources

  1. Tri Dao, personal website (tridao.me), accessed July 2026. https://tridao.me/↩↩↩↩

  2. Princeton Engineering, “Tri Dao” faculty profile, accessed July 2026. https://engineering.princeton.edu/faculty/tri-dao↩↩↩

  3. Together AI blog, “Introducing Together AI Chief Scientist Tri Dao, as he releases FlashAttention-2 to speed up model training and inference,” July 17, 2023. https://www.together.ai/blog/tri-dao-flash-attention↩↩↩↩↩

  4. Schmidt Sciences AI2050, “Tri Dao” fellow profile, accessed July 2026. https://ai2050.schmidtsciences.org/fellow/tri-dao/↩↩

  5. Tracxn, “Together AI — 2026 Company Profile, Team, Funding & Competitors,” accessed July 2026. https://tracxn.com/d/companies/together-ai/__fcIBLE0rJMeK3FAdcfzE0H41jE36bJd0FDBWalYo6cY↩↩

  6. Interconnects (Nathan Lambert), “Interviewing Tri Dao and Michael Poli on the future of LLM architectures,” December 21, 2023. https://www.interconnects.ai/p/interviewing-tri-dao-and-michael↩↩↩↩↩

  7. Data Phoenix, “Adaptive ML secured $20M in seed funding to democratize LLM preference tuning,” March 2024. https://dataphoenix.info/adaptive-ml-secured-20m-in-seed-funding/↩↩↩↩↩

  8. Sail Research blog, “Seed + Series A: Sail raises $80M in funding from Sequoia, Kleiner Perkins, and other investors,” June 24, 2026. https://www.sailresearch.com/blog/sail-raises-80m↩↩↩↩↩↩

  9. Sacra, “Prime Intellect — funding, news & analysis,” accessed July 2026 (documents February 2025 $15M seed extension led by Founders Fund with Tri Dao among individual investors). https://sacra.com/c/prime-intellect/↩↩↩↩

  10. Dedalus Labs blog, “We raised $11M to redefine how developers build AI agents,” October 15, 2025. https://www.dedaluslabs.ai/blog/dedalus-seed-round↩↩↩↩↩

  11. The AI Insider, “Sail Research Closes $80M in Funding to Build Max-Efficiency Infrastructure for AI Agents,” July 3, 2026. https://theaiinsider.tech/2026/07/03/sail-research-closes-80m-in-funding-to-build-max-efficiency-infrastructure-for-ai-agents/↩

  12. Crunchbase, “Tri Dao” person profile, accessed July 2026. https://www.crunchbase.com/person/tri-dao-eb18↩