Alex Wiltschko
Founder And CEO at Osmo
Boston, Massachusetts, United States
Summary
🤩
Rockstar🎓
Top SchoolAlex Wiltschko is a founder and CEO with 16 years of deep technical experience at the intersection of machine learning, biology, and systems engineering. Based in Boston, he led research at Google Brain and founded startups (Osmo, Syllable, Whetlab) that applied ML to sensing and preclinical biology, with multiple exits and an EIR stint at GV. He combines hands-on contributions to foundational open-source projects—JAX, Tangent, TensorFlow Probability, and low-level tooling like Terra and torch-distro—with product and build expertise in CI, packaging, and cross-platform deployment. His background spans low-latency audio systems, neural probabilistic modeling, and automatic differentiation, giving him unique fluency from kernels to models to labs. Notably, he’s worked on making complex ML tooling more reproducible and installable (Conda/Luarocks fixes, PyPI/Windows install fixes) while shipping applied sensing products that give machines new sensory capabilities.
16 years of coding experience
15 years of employment as a software developer
Bachelor of Science (B.S.), Neuroscience, Bachelor of Science (B.S.), Neuroscience at University of MichiganDoctor of Philosophy (PhD), Neuroscience, Doctor of Philosophy (PhD), Neuroscience at Harvard University
Stackoverflow
2,038reputation
507kreached
19answers
14questions
fft
top-5%
Github Skills (58)
autodiff10
pytorch10
conda-build10
gnu-make10
package-management10
operation10
scripting10
python10
tensorrt10
machine-learning10
bash10
conda10
makefile10
numpy10
lua10
Programming languages (15)
Github contributions (5)
twitter-archive/torch-autograd
May 2015 - Sep 2016
Autograd automatically differentiates native Torch code
Role in this project:
Back-end Developer
Back-end DeveloperContributions:210 commits, 45 PRs, 105 pushes in 1 year 4 months
Contributions summary:Alex primarily focused on implementing automatic differentiation capabilities for native Torch code. They developed core components for the `torch-autograd` library, including the creation of a reverse-mode automatic differentiation engine. Their work involved defining operations, gradients, and handling various tensor types to enable the computation of gradients for Torch code. They also worked on test cases for a matrix library.
pytorchautomatic-differentiationtorchmachine-learningautodiff
google/tangent
Nov 2017 - Mar 2018
Source-to-Source Debuggable Derivatives in Pure Python
Role in this project:ML Engineer
ML EngineerContributions:33 commits, 14 PRs, 30 pushes in 4 months
Contributions summary:Alex primarily contributed to the `google/tangent` repository, which focuses on automatic differentiation. Their work included fixing tests related to forward-over-reverse differentiation, which involved modifying the code in `tangent/forward_ad.py` and `tangent/utils.py`. The user also made changes related to PyPI preparation, versioning, and addressing a Windows pip install error by bumping the version, indicating involvement in the build and deployment aspects of the project. Additionally, they were involved in refactoring the API for `insert_grad_of` and added support for gradients of NumPy arrays.
pythonautomatic-differentiationdeep-learningmachine-learningautodiff
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.