Yerdos Ordabayev
Senior Machine Learning Scientist I at Pyro Probabilistic Programming Language (PPL)
Cambridge, Massachusetts, United States
Summary
🤩
Rockstar🎓
Top SchoolYerdos Ordabayev is a Senior Machine Learning Scientist with six years of experience building and deploying large-scale ML systems for biomedical data, currently driving ML efforts at Generate Biomedicines in Cambridge. He specializes in probabilistic modeling and deep generative models (Transformers, variational inference, probabilistic graphical models) and has engineered foundation-model workflows trained on trillions of tokens for single-cell modalities. A core contributor to Pyro/NumPyro/Funsor, he has extended probabilistic programming with parallel-scan inference and efficient vectorized components used in time-series and discrete-latent modeling. His work bridges research and production: he has led development of Cellarium ML and CellariumGPT, optimized multi-GPU data loaders and PCA algorithms for billion-scale datasets, and published in high-impact life-science journals. Colleagues rely on him for mentoring and for deep framework expertise (PyTorch, JAX, Pyro) that often involves modifying core APIs to unlock new performance and functionality.
6 years of coding experience
6 years of employment as a software developer
Doctor of Philosophy - PhD Computational and Molecular Biophysics, Doctor of Philosophy - PhD Computational and Molecular Biophysics at Washington University School of Medicine in St. LouisLomonosov Moscow State University
English, Kazakh, Russian, Turkish
Github Skills (21)
pytorch10
probabilistic-programming10
python10
operation10
tensorrt10
data-modeling10
machine-learning10
vectorization10
probabilistic-reasoning10
deep-learning10
tensorflow10
statistical-models10
computer-engineering10
probabilistic-models10
tensor10
Programming languages (2)
Github contributions (5)
pyro-ppl/pyro
Jan 2020 - Aug 2022
Deep universal probabilistic programming with Python and PyTorch
Role in this project:
Back-end Developer & AI/ML Engineer
Back-end Developer & AI/ML EngineerContributions:99 reviews, 91 commits, 76 PRs in 2 years 6 months
Contributions summary:Yerdos contributed to the Pyro probabilistic programming library by addressing bug fixes and improving code efficiency. Their work included fixing issues in the `MaskedMixture` distribution and expanding the functionality of the `MaskedDistribution`. They also introduced a vectorized Markov Messenger, enhancing the library's capabilities for time-series modeling and inference. Furthermore, the user improved documentation and addressed typos within the library.
pytorchpythondeep-learningbayesian-inferenceprobabilistic-modeling
cellarium-ai/cellarium-ml
Jan 2023 - Mar 2025
Distributed single-cell data analysis.
Contributions:3 releases, 135 reviews, 249 PRs in 2 years 2 months
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.