Emily Fertig
Software Engineer at Google
San Francisco Bay Area United States
Summary
🤩
Rockstar🎓
Top SchoolEmily Fertig is a software engineer with six years of experience at Google focused on numerical computing and ML infrastructure, now working on JAX. She brings deep expertise in TensorFlow and TensorFlow Probability—authoring gradient support for LinearOperator classes and improving probabilistic distributions—and has contributed core lowering rules and testing to the JAX ecosystem. Her background spans research and applied roles from AI Residency to Research Software Engineer, blending probabilistic ML, Bayesian methods, and compiler-level improvements for XLA. Prior work in energy policy and academic research (PhD-level training) gives her a rare cross-disciplinary perspective on rigorous experimentation and system-level tradeoffs. Based in the San Francisco Bay Area, she is an active open-source contributor to flagship projects like TensorFlow and JAX, improving both correctness and performance. Colleagues rely on her ability to translate mathematical abstractions into robust, tested code that scales to production ML workloads.
6 years of coding experience
12 years of employment as a software developer
Sidwell Friends SchoolPhD, Engineering and Public Policy, PhD, Engineering and Public Policy at Carnegie Mellon UniversityBA, Geosciences, BA, Geosciences at Williams College
Russian, Spanish, Norwegian
Stackoverflow
1reputation
0reached
0answers
0questions
Github Skills (31)
probabilistic-programming10
python10
operation10
tensorrt10
data-modeling10
machine-learning10
probabilistic-reasoning10
deep-learning10
tensorflow10
statistical-models10
optmization10
resnet10
optimisation10
probabilistic-models10
tensor10
Programming languages (5)
Github contributions (5)
tensorflow/probability
Aug 2019 - Jan 2023
Probabilistic reasoning and statistical analysis in TensorFlow
Role in this project:
ML Engineer & Data Scientist
ML Engineer & Data ScientistContributions:10 releases, 8 reviews, 271 commits in 3 years 4 months
Contributions summary:Emily's commits focused on adding the WeightNorm wrapper for weight normalization of layers in the `tensorflow/probability` repository, aligning with the project's deep learning focus. They also contributed significantly to the mathematical foundations of various distributions by deprecating an existing full covariance distribution and providing a replacement using the more efficient MultivariateNormalTriL distribution. Additional contributions included refining docstrings for existing distributions to improve their usability.
statisticspythonprobabilistic-reasoningdata-sciencedeep-learning
jax-ml/jax
Jul 2022 - Mar 2025
Composable transformations of Python+NumPy programs: differentiate, vectorize, JIT to GPU/TPU, and more
Role in this project:ML Engineer
ML EngineerContributions:12 reviews, 13 PRs, 32 comments in 2 years 7 months
Contributions summary:Emily primarily contributed to the development of Mosaic, a component within the JAX ecosystem, by implementing lowering rules for various mathematical operations. These changes include adding support for functions like `absf`, `fpowi`, `log1p`, `min`, `max`, `sin`, and `sqrt`. The user also improved error messages and added a testing framework, contributing to the overall robustness and functionality of the Mosaic framework. Furthermore, the user added support for multiple branches in `cond`.
pytorchpythonjitautomatic-differentiationgpu
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.