PeopleofSuperintelligence

Builder

Jianmin Chen

Jianmin Chen is a software engineer at Google and a co-author of the 2016 paper that describes TensorFlow, the open-source machine learning system from the Google Brain team. They studied for a computer science PhD at the University of Florida from 2009 to 2014, publishing work on GPU execution ordering, random-forest models of GPU power and GPU scheduling. After joining Google in Santa Clara in 2014, they worked on the TensorFlow training libraries, which Google thanked them for when it released TensorFlow 0.8 with distributed computing support in April 2016. In 2016, with Rajat Monga, Samy Bengio and Rafal Jozefowicz, they showed that synchronous SGD with backup workers can train deep networks to better accuracy in less wall-clock time.

Now at
Google
Badges

Timeline

  1. 2009–2014

    Computer science PhD at Florida

    They study for a computer science PhD at the University of Florida in Gainesville from 2009 to 2014. Their research covers computer architecture, GPGPU and parallel computing.

  2. 2010

    Weak execution ordering on GPUs

    Working in the University of Florida's Department of Computer and Information Science and Engineering, they present "Weak execution ordering" at the 2010 IEEE International Symposium on Performance Analysis of Systems and Software in March 2010, with Jih-Kwon Peir, Lu Peng and three other co-authors. Their method overlaps computation with communication between GPU thread blocks and speeds up Poisson image editing and shape from shading 4 to 5 times on a Tesla C1060, compared with host synchronization.

  3. 2011–2013

    GPU power models and scheduling

    In 2011, they model GPU power with tree-based random forest methods, using CUDA kernels measured on a GTX 280, and co-author a comparison of Nvidia's Fermi and ATI's Cypress GPU architectures. In 2013, they publish "Guided region-based GPU scheduling", which uses multi-thread parallelism to hide memory latency.

  4. 2014

    Joining Google in Santa Clara

    They join Google in Santa Clara as a software engineer in 2014 and contribute to the TensorFlow training libraries.

  5. April 2016

    Distributed training in TensorFlow 0.8

    Google releases TensorFlow 0.8 with distributed computing support in April 2016, so that models can train on hundreds of machines in parallel. Google thanks them, with Matthieu Devin, Sherry Moore and Sergio Guadarrama, for the TensorFlow training libraries in the release.

  6. 2016

    Synchronous SGD with backup workers

    With Rajat Monga, Samy Bengio and Rafal Jozefowicz, they present "Revisiting Distributed Synchronous SGD" in the ICLR 2016 workshop track. The paper shows that synchronous training with backup workers avoids asynchronous noise, limits the effect of the slowest workers, and reaches better accuracy in less wall-clock time. Its arXiv version, first posted in April 2016, also lists Xinghao Pan as a co-author.

  7. November 2016

    The TensorFlow system paper

    They co-author "TensorFlow: A System for Large-Scale Machine Learning" with Martín Abadi, Jeffrey Dean and 19 other Google Brain colleagues, presented at OSDI in Savannah, Georgia, in November 2016. The paper describes how TensorFlow maps dataflow graphs across the machines of a cluster and across CPUs, GPUs and Tensor Processing Units.

  8. As of October 2026

    Software engineer at Google

    They work as a software engineer at Google, the company they joined in 2014.

Career

  1. Since 2014GoogleSoftware engineer

Education

  1. 2014University of FloridaPhD

From the biography.

  1. Google had the most people shaping AI every year from 2014 to 2025Includes Google. Our estimates of which organizations employed the most people shaping AI each year from 1955 to 2025, tracing the shift from university labs to companies.
  2. Google has the highest combined h-index among AI organizationsIncludes Google. We estimate that Google holds the highest combined h-index at 8,016, while universities account for 73% of the deduplicated total across AI organizations.
  3. 20% of people at companies who shape AI also hold a university role, led by Google and Stanford UniversityIncludes Google. We estimate that 539 of the 6,766 people with a known current organization (8%) hold both a university role and a company role, including 20% of people at companies and 13% of people at universities.
  4. Google alumni have founded or led 133 other organizations, more than any other AI labIncludes Google. We estimate that alumni of 17 major AI labs went on to found or lead 350 other organizations, including 226 they founded, and we trace 472 career moves directly between 23 labs since 2010.

Colleagues

People named in this biography, and people with dated positions at the same organization in the same years.

Sources and credits

Photo: We have no published photo of Jianmin Chen yet. Send us a free-licensed photo.

  1. jianmin chen | alphaXiv alphaxiv.org. Checked 20 October 2026.
  2. jianmin chen - Google Scholar scholar.google.com. Checked 20 October 2026.
  3. Weak execution ordering – exploiting iterative methods on many-core GPUs | hgpu.org hgpu.org. Checked 20 October 2026.
  4. "Statistical GPU power analysis using tree-based methods" by Jianmin Chen, Bin Li et al. repository.lsu.edu. Checked 20 October 2026.
  5. "Architecture comparisons between Nvidia and ATI GPUs: Computation parallelism and data communications" by Ying Zhang, Lu Peng et al. repository.lsu.edu. Checked 20 October 2026.
  6. Announcing TensorFlow 0.8 – now with distributed computing support! research.google. Checked 20 October 2026.
  7. Revisiting Distributed Synchronous SGD research.google. Checked 20 October 2026.
  8. [1604.00981] Revisiting Distributed Synchronous SGD arxiv.org. Checked 20 October 2026.
  9. TensorFlow: A System for Large-Scale Machine Learning | USENIX usenix.org. Checked 20 October 2026.
  10. TensorFlow: A system for large-scale machine learning research.google. Checked 20 October 2026.