DZ
DefZone.Net

Kien thuc cong nghe & giai tri

  1. Home
  2. Wiki
For the punk rock musician, see Jeff Dean (musician).

Jeff Dean
Dean in 2025
Born (1968-07-23) July 23, 1968 (age 57)
Hawaii
Alma materUniversity of Minnesota, B.S. Computer Science and Engineering (1990)
University of Washington, Ph.D. Computer Science (1996)
Known forMapReduce, Bigtable, Spanner, TensorFlow
Scientific career
FieldsComputer science
InstitutionsGoogle, Digital Equipment Corporation
ThesisWhole-program optimization of object-oriented languages (1996)
Doctoral advisorCraig Chambers

Jeffrey Adgate Dean (born July 23, 1968) is an American computer scientist and software engineer. Since 2018, he has been the lead of Google AI.[1] He was appointed Google's chief scientist in 2023 after the merger of DeepMind and Google Brain into Google DeepMind.[2]

Early life and education

[edit]

Jeff Dean was born to a family of a tropical disease researcher and a medical anthropologist, who moved the family frequently. However, from the fifth to tenth grade, Dean attended schools in the Twin Cities.[3]

Dean received a B.S., summa cum laude, from the University of Minnesota in computer science and economics in 1990.[4] His undergraduate thesis was on neural networks in C programming, advised by Vipin Kumar.[5][6] During freshman year, Dean met his future wife, Heidi Hopper, who graduated from the same university in 1990.[3]

He received a Ph.D. in computer science from the University of Washington in 1996, working under Craig Chambers on compilers[7] and whole-program optimization techniques for object-oriented programming languages.[8]

Career

[edit]

Early career

[edit]

Before graduate school, Dean worked at the World Health Organization's Global Programme on AIDS, developing software for statistical modeling and forecasting of the HIV/AIDS pandemic.[9]

After graduate school, Dean worked at DEC/Compaq's Western Research Laboratory,[10] on profiling tools, microprocessor architecture and information retrieval.[9] Much of his work was completed in close collaboration with Sanjay Ghemawat.[10][7] In early 1999 he briefly worked at comparison-shopping startup mySimon as a senior member of technical staff, designing and implementing a distributed system for web crawling, caching, and full-text indexing of product information to improve the scalability of its shopping search service, before leaving later that year to join Google.

Career at Google

[edit]

Dean joined Google in mid-1999.[4] He was Google's 30th employee.[11] As of 2018, Dean and Sanjay Ghemawat are the only two employees at Google to hold the title of Senior Fellow, the highest technical level at the company.[7]

Systems design

[edit]

In collaboration with Sanjay Ghemawat, Dean designed and implemented portions of Google's advertising, crawling, indexing and query serving systems, along with components of the distributed computing infrastructure that underlies most of Google's products.[7] He has also worked on improving search quality, statistical machine translation (Google translate) and internal software development tools and has had significant involvement in the engineering hiring process.

Dean and Ghemawat led the original design of Protocol Buffers, an open-source data interchange format that provides language-neutral and platform-neutral mechanisms for serializing structured data.[12] Protocol Buffers are used extensively across Google for remote procedure call (RPC) protocols and for storing structured information in persistent storage systems, offering advantages in efficiency and extensibility compared to alternatives like XML.[13]

The pair also developed MapReduce, a programming model and implementation for processing and generating large datasets that became foundational for applications.[7] MapReduce abstracts away the complexities of distributed computing (including data partitioning, inter-machine communication, and fault handling) allowing programmers without specialized distributed systems knowledge to leverage clusters of thousands of commodity machines for data processing tasks. The system's success at Google subsequently inspired open-source implementations including Apache Hadoop, which democratized large-scale data processing.

Bigtable, co-designed by Dean and Ghemawat alongside other Google engineers including Fay Chang and Mike Burrows, represents a large-scale semi-structured storage system designed to handle the petabyte-scale data requirements of Google's products.[7] Bigtable is engineered to store sparsely populated tables that scale to billions of rows and thousands of columns while maintaining low-latency, high-throughput read and write operations, and it integrates seamlessly with MapReduce for batch processing.[14] The system's architecture and design influenced the NoSQL database movement, with multiple open-source projects adopting similar approaches.

Spanner represents Dean and Ghemawat's contribution to globally distributed database systems, delivering a horizontally scalable, multi-version database that can replicate data synchronously across geographically distributed datacenters while maintaining strong consistency guarantees.[15] Spanner's architecture enables applications to specify fine-grained replication constraints, control data locality to minimize latency, and transparently migrate data between datacenters for load balancing, supporting SQL-based queries and ACID transactions across potentially millions of machines and trillions of rows.

Dean and Ghemawat developed LevelDB, an open-source on-disk key-value store released in 2011 that drew inspiration from Bigtable's design principles while minimizing external dependencies to enable broader adoption.[16] LevelDB stores sorted key-value pairs, supports atomic batch writes, forward and backward iteration, and compression via Google's Snappy library, subsequently becoming the backend storage engine for Google Chrome's IndexedDB as well as other applications including Bitcoin Core and Minecraft Bedrock Edition.[17]

Artificial intelligence

[edit]

Dean joined Google X in 2011 to investigate deep neural networks, which had just resurged in popularity. This ended with "the cat neuron paper", a deep belief network trained by unsupervised learning on YouTube videos.[18] This project morphed into Google Brain, a team that studies large-scale artificial neural networks, also formed in 2011.[7] Jeff Dean became its leader in 2012.

Dean worked on DistBelief, a proprietary machine-learning system for distributed training of deep neural networks, whose name reflects the possibility to train deep belief networks.​ DistBelief was eventually refactored into TensorFlow, an open-source machine-learning software library, for which he was among the designers and implementers of the initial release.[13][7]​ The DistBelief system was used to train the network in "the cat neuron paper".[18][19] TensorFlow initially dominated the machine-learning research landscape, but by 2023 accounted for only 8% of exclusive models on Hugging Face compared to 92% for PyTorch, which was also used in 80% of academic papers utilizing deep learning frameworks.[20]

Dean also worked on Pathways, an asynchronous distributed dataflow system for neural networks.[21][13] It was used in PaLM.[22]

Dean in 2025

In April 2018, Dean was appointed the head of Google's artificial intelligence division, after John Giannandrea left to lead Apple's AI projects.[23][24]

Main article: Criticism_of_Google § Departures of researchers

In December 2020, Google’s Ethical AI team co-lead Timnit Gebru was terminated after a draft paper she co-authored on the risks and ethical implications of large language models didn't receive publication approval. The incident received significant attention in the media, and in the aftermath, Dean addressed an email to staff acknowledging that the episode had “surfaced large, important issues” around research culture, bias, and inclusion within Google’s AI organization.[25][26][27][28] In an open letter, critics demanded that members of Google's senior leadership, including Dean, "explain the process by which the paper was unilaterally rejected by leadership."[29] Later, Margaret Mitchell, then co-lead of the Ethical AI team and an outspoken supporter of Gebru, was fired on February 19, 2021 for "reportedly using automated scripts to find examples of mistreatment of Dr. Timnit Gebru."[30] On the same day Mitchell was fired, Dean issued an internal memo acknowledging Google "could have and should have handled this situation with more sensitivity" and said he "regrets that it led to some employees questioning whether they belong at Google."[31] The media reported that Google's reputation with AI researchers suffered as a result.[32]

Main article: Criticism_of_Google § AlphaChip controversy

In June 2021, Dean was a senior author on a Nature paper introducing a reinforcement-learning approach to chip floorplanning. Dean and co-authors claimed the method outperformed prior approaches.[33] Google engineer Satrajit Chatterjee raised concerns about the results to Dean.[34][35] Google denied Chatterjee permission to publish his critique and terminated him.[36][35] The Satrajit Chatterjee v. Google LLC lawsuit named Dean and alleged fraud, retaliation for Chatterjee's critique, and wrongful termination.[35] A California court denied Google's motion to dismiss Chatterjee's lawsuit.[37] Academic researchers raised criticisms of the Nature paper's results and disclosure of methods.[34][36][38] In 2023, Professor Andrew Kahng and UCSD collaborators reported that replication studies on public benchmarks found simulated annealing and commercial EDA software to outperform the Nature method.[39][40]

In 2023, DeepMind was merged with Google Brain to form a unified AI research unit, Google DeepMind, headed by Demis Hassabis. As part of this reorganization, Dean became Google's chief scientist.[2][13] Dean proposed the name Gemini for the chatbot developed by Google DeepMind, "because it's like twins coming together".[11]

Philanthropy

[edit]

Dean and his wife, Heidi Hopper, started the Hopper-Dean Foundation and began making philanthropic grants in 2011. In 2016, the foundation gave $2 million each to UC Berkeley, Massachusetts Institute of Technology, University of Washington, Stanford University and Carnegie Mellon University to support programs that promote diversity in science, technology, engineering and mathematics (STEM).[41] The Menlo Park-based foundation gave $22.1 million to a variety of universities and non-profits in 2023 and ended the year with $54.4 million in assets, according to its Form 990.[42]

In 2025, Dean joined the board of Laude Institute, steering the organization with David Patterson, Joelle Pineau, and Andy Konwinski.[43]

Awards and honors

[edit]

Publications

[edit]

Selected papers

[edit]

Personal life

[edit]

Dean is married and has two daughters.[7]

He is the subject of an Internet meme for "Jeff Dean facts". Similar to Chuck Norris facts, the Jeff Dean facts exaggerate his programming powers.[50] For example:[51]

Once, in early 2002, when the index servers went down, Jeff Dean answered user queries manually for two hours. Evals showed a quality improvement of 5 points.

See also

[edit]

References

[edit]
  1. ^ Vincent, James (April 3, 2018). "Google veteran Jeff Dean takes over as company's AI chief". The Verge. Retrieved November 29, 2023.
  2. ^ a b Elias, Jennifer (April 20, 2023). "Read the internal memo Alphabet sent in merging A.I.-focused groups DeepMind and Google Brain". CNBC. Retrieved June 22, 2023.
  3. ^ a b Gehler, Chris (2013). "Jeff Dean: Google's unsung hero". College of Science and Engineering. University of Minnesota. Retrieved December 6, 2025.
  4. ^ a b "Jeff Dean". research.google.
  5. ^ Dean, Jeffrey (1990). Parallel implementations of neural network training: Two back-propagation approaches (Thesis). University of Minnesota.
  6. ^ @jeffdean (August 27, 2018). "[...] Kudos to University of Minnesota (@UMNews) Honors Program. Earlier this year, I asked Prof. Vipin Kumar, my advisor for this work, if he still had a copy, since I had lost my copy. He didn't, but checked with the Honors Program [...]" (Tweet) – via Twitter.
  7. ^ a b c d e f g h i Somers, James (December 3, 2018). "The Friendship That Made Google Huge". The New Yorker. ISSN 0028-792X. Retrieved September 28, 2025.
  8. ^ "STANFORD TALKS; Jeff Dean: TensorFlow Overview and Future Directions". Stanford University. January 21, 2016. Archived from the original on August 28, 2016. Retrieved August 15, 2016.
  9. ^ a b "Jeff Dean". Speakerpedia. Retrieved August 19, 2016.
  10. ^ a b Metz, Cade (August 8, 2012). "If Xerox PARC Invented the PC, Google Invented the Internet". Wired. Retrieved December 16, 2017.
  11. ^ a b "Time100 AI 2025 | Jeffrey Dean". Time. 2025.
  12. ^ "Google shares more of its secret sauce: Protocol buffers". ZDNET. July 8, 2008. Retrieved November 25, 2025.
  13. ^ a b c d "Jeffrey Dean". Google Research. Retrieved October 25, 2024.
  14. ^ "Bigtable overview". Google Cloud. Retrieved December 20, 2025.
  15. ^ "Spanner: Google's Globally-Distributed Database". December 31, 2022.
  16. ^ "Google Open Source Blog: LevelDB: A Fast Persistent Key-Value Store". Google, Inc.
  17. ^ "google/leveldb". Retrieved December 20, 2025.
  18. ^ a b Le, Quoc V. (May 2013). "Building high-level features using large scale unsupervised learning". 2013 IEEE International Conference on Acoustics, Speech and Signal Processing. IEEE. pp. 8595–8598. arXiv:1112.6209. doi:10.1109/icassp.2013.6639343. ISBN 978-1-4799-0356-6.
  19. ^ Markoff, John (June 25, 2012). "How Many Computers to Identify a Cat? 16,000". The New York Times.
  20. ^ "PyTorch vs TensorFlow in 2023". AssemblyAI. Retrieved December 6, 2025.
  21. ^ "Google to build more efficient, multi-capability AI systems". The Register. March 31, 2022.
  22. ^ "PaLM: Scaling Language Modeling with Pathways" (PDF). JMLR. 2023.
  23. ^ Simonite, Tom. "Google's New AI Head Is So Smart He Doesn't Need AI". Wired. ISSN 1059-1028. Retrieved November 29, 2024.
  24. ^ D'Onfro, Jillian (April 2, 2018). "Google is splitting A.I. into its own business unit and shaking up its search leadership". CNBC. Retrieved April 3, 2018.
  25. ^ Hao, Karen (December 4, 2020). "Google pushed out an AI ethics researcher days after she published a paper on risks of language models". MIT Technology Review. Retrieved October 1, 2025.
  26. ^ Garrahan, Matthew (December 4, 2020). "Google's AI ethics investigator Timnit Gebru was forced out after a paper on big tech risks". Vox. Retrieved October 1, 2025.
  27. ^ Banjo, Shelly; Bergen, Mark (December 10, 2020). "The withering email that got an ethical AI researcher fired at Google". Platformer. Retrieved October 1, 2025.
  28. ^ Langley, Hannah (December 15, 2020). "Google's AI chief canceled a year-end gathering in light of the backlash over researcher Timnit Gebru's exit: 'A celebration doesn't seem appropriate at this time'". Business Insider. Retrieved October 1, 2025.
  29. ^ Allyn, Bobby; Metz, Rachel (December 16, 2020). "Google workers demand apology and reinstatement for ousted AI researcher Timnit Gebru". The Guardian.
  30. ^ Dickey, Megan Rose (February 19, 2021). "Google fires top AI ethics researcher Margaret". TechCrunch.
  31. ^ Rachel Metz (March 11, 2021). "How one employee's exit shook Google and the AI industry". CNN.
  32. ^ James Vincent (April 13, 2021). "Google is poisoning its reputation with AI researchers". The Verge. Retrieved August 1, 2025.
  33. ^ Mirhoseini, Azalia; Goldie, Anna; Yazgan, Mustafa; Jiang, Joe Wenjie; Songhori, Ebrahim; Wang, Shen; Lee, Young-Joon; Johnson, Eric; Pathak, Omkar; Nazi, Azade; Pak, Jiwoo; Tong, Andy; Srinivasa, Kavya; Hang, William; Tuncer, Emre; Le, Quoc V.; Laudon, James; Ho, Richard; Carpenter, Roger; Dean, Jeff (2021). "A graph placement methodology for fast chip design". Nature. 594 (7862): 207–212. Bibcode:2021Natur.594..207M. doi:10.1038/s41586-021-03544-w. PMID 34108699.
  34. ^ a b Dave, Paresh (May 3, 2022). "Google faces internal battle over research on AI to speed chip design". Reuters. Retrieved December 26, 2025.
  35. ^ a b c Satrajit Chatterjee v. Google, LLC (Case no. 22CV398683) (Superior Court of California, County of Santa Clara February 21, 2023) ("The full complaint as filed in Santa Clara County, detailing Chatterjee’s whistleblower allegations and claims against Google."), Text.
  36. ^ a b Wakabayashi, Daisuke; Metz, Cade (May 2, 2022). "Another firing among Google's A.I. brain trust, and more discord". The New York Times. Retrieved December 26, 2025.
  37. ^ Burnson, Robert. "Fired Google AI Engineer's Whistleblower Lawsuit Moves Ahead". Bloomberg Law. Retrieved September 3, 2025.
  38. ^ Goth, Gregory (March 29, 2023). "More Details, but Not Enough". Communications of the ACM. Retrieved June 13, 2024.
  39. ^ "Google's claims of super-human AI chip layout in dispute". The Register. March 27, 2023. Retrieved August 5, 2025.
  40. ^ Cheng, Chung-Kuan; Kahng, Andrew B.; Kundu, Sayak; Wang, Yucheng; Wang, Zhiang (2023). "Assessment of Reinforcement Learning for Macro Placement". Proceedings of the 2023 International Symposium on Physical Design (ISPD ’23). ACM. pp. 158–166. doi:10.1145/3569052.3578926.
  41. ^ "$1M Hopper-Dean Foundation Gift for Diversity in CS". UC Berkeley. Retrieved January 29, 2019.
    - Williams, Tate (August 10, 2016). "One of Google's Top Programmers Has Made STEM Diversity a Philanthropic Cause". Inside Philanthropy. Retrieved August 15, 2016.
    - "$1 million gift to support diversity in STEM education". Massachusetts Institute of Technology. Retrieved October 25, 2020.
  42. ^ "HOPPER-DEAN FOUNDATION C/O CATALYST FAMILY OFFICE LLC — 2023 990PF". irs-efile-renderer.instrumentl.com. Retrieved September 17, 2025.
  43. ^ Sullivan, Mark (June 23, 2025). "This Perplexity cofounder wants to help AI breakthroughs graduate from university labs". Fast Company. Retrieved September 30, 2025.
  44. ^ "The Mark Weiser Award". ACM SIGOPS. Retrieved July 5, 2019.
  45. ^ "Jeff Dean elected to National Academy of Engineering". UW CSE News. University of Washington. February 5, 2009. Retrieved August 15, 2016.
    - "Jeffrey A Dean - Award Winner". Association for Computing Machinery. Retrieved August 15, 2018.
  46. ^ a b "Jeffrey A Dean". ACM Awards Home.
  47. ^ "The Mark Weiser Award". sigops.org.
  48. ^ Newly Elected Members, American Academy of Arts and Sciences, April 2016, retrieved April 20, 2016
  49. ^ "IEEE John Von Neumann Recipients" (PDF). corporate-awards.ieee.org.
  50. ^ Carlson, Nicholas. "Astounding 'Facts' About Google's Most Badass Engineer, Jeff Dean". Business Insider. Retrieved November 30, 2024.
  51. ^ Ritzdorf, Lucas (October 23, 2024), LRitzdorf/TheJeffDeanFacts, retrieved November 29, 2024
[edit]