PeopleofSuperintelligence

Researcher

James Bergstra

James Bergstra is a machine-learning researcher and software developer whose work on Theano, random search, and Hyperopt shaped how neural networks are compiled and tuned. At the University of Montreal, he completed a master's degree under Douglas Eck in 2006, served as the first lead author of the Theano deep learning library, and finished a PhD under Yoshua Bengio in July 2011. With Bengio, he published a 2012 paper on random search for hyperparameter optimization, and his postdoctoral research with David Cox at Harvard produced the Hyperopt library. He later cofounded Kindred Systems in 2014, served as Director of AI Platform at Ocado Technology, and in 2026 leads the PlanZero climate-modeling project.

Now at
Kindred, Ocado Technology and Creative Destruction Lab
Badges

Timeline

  1. 2006–July 2011

    Theano and a Montreal doctorate

    In 2006, he completes a master's degree under Douglas Eck on algorithms for classifying recorded music by genre. As the first lead author of the Theano deep learning library, he publishes "Theano: A CPU and GPU Math Compiler in Python" in the June 2010 SciPy conference proceedings. Bergstra finishes his PhD in deep learning at the University of Montreal under Yoshua Bengio in July 2011 with a dissertation on incorporating complex cells into deep learning models.

  2. 2011–2012

    Random search for hyperparameter optimization

    With Rémi Bardenet, Yoshua Bengio, and Balázs Kégl, he coauthors "Algorithms for Hyper-parameter Optimization" for Neural Information Processing Systems in 2011. In 2012, he and Bengio publish "Random Search for Hyper-Parameter Optimization" in the Journal of Machine Learning Research, showing empirically and theoretically that randomly chosen trials are more efficient for hyperparameter optimization than trials on a grid.

  3. September 2012

    Banting fellowship at Waterloo

    In September 2012, he receives a Banting Postdoctoral Fellowship from the Natural Sciences and Engineering Research Council of Canada. Moving from the Rowland Institute at Harvard, he joins the Centre for Theoretical Neuroscience at the University of Waterloo in September 2012 to work with Chris Eliasmith.

  4. 2013–2015

    Model search and Hyperopt

    Through postdoctoral work at Harvard with David Cox, he develops hyperparameter optimization methods that become the Hyperopt project. In 2013, he coauthors "Making a Science of Model Search" with Daniel Yamins and Cox at the 30th International Conference on Machine Learning, proposing a meta-modeling approach to automated hyperparameter optimization. A 2015 paper presents the Hyperopt library, which provides algorithms and parallelization infrastructure for model selection in Python.

  5. May 2016

    The Theano framework paper

    In May 2016, he is among the authors of the Theano Development Team paper on Theano, a Python library for defining, optimizing, and evaluating mathematical expressions on multi-dimensional arrays. The paper notes that development of the library began in 2008.

  6. September 2018

    Benchmarking reinforcement learning on robots

    In September 2018, he coauthors "Benchmarking Reinforcement Learning Algorithms on Real-World Robots". The study introduces reinforcement learning tasks with multiple commercially available robots and tests four off-the-shelf algorithms and their hyperparameter sensitivity on those tasks.

  7. March 2020

    Kindred Systems and Ocado Technology

    After cofounding the robotics company Kindred Systems in 2014, he serves as its Chief Science Officer, Chief Technology Officer, and Head of AI Research, presenting at the MODEX Emerging Technologies Theater in March 2020. He also works as Director of AI Platform at Ocado Technology and as a scientist in the Artificial Intelligence stream at CDL-Toronto.

  8. December 2025–April 2026

    Climate modeling with PlanZero

    In December 2025, he launches PlanZero to quantify how climate strategies affect economics, emissions, and accumulated heat. As the project's lead author and developer in 2026, he publishes an April 2026 analysis modeling the emissions effects of Canadian beef and dairy farmers adopting the feed additive Bovaer.

Career

  1. NowKindredCo-founder, Chief Science Officer
  2. NowOcado TechnologyDirector of AI Platform
  3. NowCreative Destruction LabScientist
  4. Harvard UniversityPostdoctoral researcher

Education

  1. 2010Université de MontréalPhD

From the biography.

  1. More people shaping AI studied at Massachusetts Institute of Technology than at any other institutionIncludes Harvard University. We estimate where the people shaping AI studied and where their careers led. Massachusetts Institute of Technology leads with 467 former students, including 291 recorded doctorates.
  2. The median h-index is 48 at universities and 14 at frontier AI labsIncludes Harvard University. The median h-index is 48 at universities and 14 at frontier AI labs. We compare publishing careers, recent citations, dual appointments and job moves across 2,420 university researchers and 308 frontier-lab researchers.
  3. Companies supply the most prior-work hires at 7 of the 9 frontier AI labsIncludes Harvard University. Our estimates of the organizations where 550 people worked directly before joining 9 frontier AI labs across 620 career transitions, comparing the main feeder organizations per lab.
  4. 44% of AI career moves go from one university to anotherIncludes Harvard University. We estimate how 26,439 AI career moves link universities, companies, labs, nonprofits and government research centers, with decade counts, net flows and the top paths.

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 James Bergstra yet. Send us a free-licensed photo.

  1. CNRGlab @ UWaterloo | People compneuro.uwaterloo.ca. Checked 22 October 2026.
  2. James Bergstra - Creative Destruction Lab creativedestructionlab.com. Checked 22 October 2026.
  3. Theano: A CPU and GPU Math Compiler in Python - SciPy Proceedings proceedings.scipy.org. Checked 22 October 2026.
  4. J. Bergstra Publications eng.uwaterloo.ca. Checked 22 October 2026.
  5. Random Search for Hyper-Parameter Optimization jmlr.org. Checked 22 October 2026.
  6. Prestigious Banting Postdoctoral Fellows come to Waterloo | Waterloo News | University of Waterloo uwaterloo.ca. Checked 22 October 2026.
  7. Making a Science of Model Search: Hyperparameter Optimization in Hundreds of Dimensions for Vision Architectures proceedings.mlr.press. Checked 22 October 2026.
  8. CNRGlab @ UWaterloo | Publications compneuro.uwaterloo.ca. Checked 22 October 2026.
  9. Theano: A Python framework for fast computation of mathematical expressions arxiv.org. Checked 22 October 2026.
  10. Benchmarking Reinforcement Learning Algorithms on Real-World Robots arxiv.org. Checked 22 October 2026.
  11. Robo Crop: Seven B.C. companies pushing the boundaries of robotics - BCBusiness bcbusiness.ca. Checked 22 October 2026.
  12. New Kindred SORT 3.7 Drives Throughput for E-commerce Fulfillment with Expanded Putwall and Improved Material Handling 2020.modexshow.com. Checked 22 October 2026.
  13. PlanZero planzero.ca. Checked 22 October 2026.
  14. PlanZero planzero.ca. Checked 22 October 2026.
  15. PlanZero planzero.ca. Checked 22 October 2026.