John Salvatier
Researcher at AI Impacts
Berkeley, California, United States
Summary
🤩
Rockstar🎓
Top SchoolJohn Salvatier is a researcher and software engineer with 17 years of experience bridging Bayesian statistics, scientific computing, and production-scale engineering. Based in Berkeley, he is the primary author of PyMC3 and has contributed core fixes and features to foundational scientific projects like NumPy and Theano, reflecting deep expertise in probabilistic programming and numerical libraries. His industry background includes building functional, large-scale data systems at Amazon and quantitative trading software for RPX Research, giving him both research rigor and production sensibility. He helped grow the Seattle Effective Altruists community from its inception, signaling a long-standing commitment to applied ethics and community building alongside his technical work. Notably, his open-source contributions often focus on correctness and stability—bug fixes, gradient and broadcasting fixes, and targeted tests—that have improved trustworthiness in widely used scientific code.
17 years of coding experience
6 years of employment as a software developer
B.S.Ch.E., B.S., Chemical Engineering, Paper Science and Engineering, B.S.Ch.E., B.S., Chemical Engineering, Paper Science and Engineering at University of Washington
Stackoverflow
3,147reputation
268kreached
27answers
29questions
Github Skills (23)
python10
operation10
testing10
tensorrt10
distributions10
gradient10
numpy10
tensorflow10
statistical-models10
tensor10
probability-distribution10
linear-algebra10
numerical-computing10
theano10
c119
Programming languages (8)
Github contributions (5)
pymc-devs/pymc
May 2010 - Dec 2016
Bayesian Modeling and Probabilistic Programming in Python
Role in this project:
Back-end Developer
Back-end DeveloperContributions:717 commits, 58 PRs, 85 pushes in 6 years 8 months
Contributions summary:John's commits focus on implementing and fixing issues in the PyMC library, particularly related to statistical distributions. The changes include fixing bugs in the negative binomial distribution and adding tests for NumPy deterministics. The user also added the lognormal distribution and handled issues with the Dirichlet distribution, demonstrating a strong focus on improving the core functionality and statistical models within the library.
pythonbayesian-inferencestatistical-inferencemachine-learningprobabilistic-programming
Theano/Theano
Jun 2012 - Jun 2015
Theano was a Python library that allows you to define, optimize, and evaluate mathematical expressions involving multi-dimensional arrays efficiently. It is being continued as PyTensor: www.github.com/pymc-devs/pytensor
Role in this project:Back-end Developer
Back-end DeveloperContributions:43 commits, 1 PR, 3 comments in 2 years 11 months
Contributions summary:John primarily focused on improving the functionality and stability of the Theano library. Their contributions involved fixing bugs related to broadcasting and gradients in core tensor operations. They made enhancements to advanced indexing, including improvements to the `AdvancedSubtensor` and `AdvancedIncSubtensor` operations to ensure correct behavior. Further, the user worked to incorporate newer versions of NumPy and enhance support for various data types within the library's core computational elements.
python-librarymathmulti-dimensionalpythonevaluate
Find and Hire Top DevelopersWe’ve analyzed the programming source code of over 60 million software developers on GitHub and scored them by 50,000 skills. Sign-up on Prog,AI to search for software developers.