Justas Dauparas
Co-Founder at Xaira Therapeutics
Seattle, Washington, United States
Summary
🤩
Rockstar🎓
Top SchoolJustas Dauparas is a computational biophysicist and co-founder with nine years of experience applying math, physics and modern ML to biological problems. He completed a PhD in Applied Mathematics at Cambridge and has held postdoctoral positions at Harvard and the Institute for Protein Design, where he built variational autoencoders and other models for protein design and single-cell RNA analysis. His work spans self-supervised and reinforcement learning — including policy-gradient and deep Q-learning experiments at Microsoft Research — and practical MLOps contributions to high-impact open-source tools like ProteinMPNN. Based in Seattle, he combines deep theoretical training with hands-on engineering: improving model usability, experiment reproducibility, and CA-only protein models. As a founder at Xaira Therapeutics he’s translating research-grade methods into translational biotech products, bringing an unusual blend of rigorous modelling and production-focused implementation.
9 years of coding experience
Doctor of Philosophy - PhD, Applied Mathematics, Doctor of Philosophy - PhD, Applied Mathematics at University of CambridgePanevėžys Juozas Balčikonis gymnasium
Lithuanian, English, German
Github Skills (5)
pytorch10
machine-learning10
python10
ml-deployment9
continuous-deployment9
Programming languages (2)
Github contributions (5)
dauparas/ProteinMPNN
May 2022 - Nov 2022
Code for the ProteinMPNN paper
Role in this project:
Back-end Developer & MLOps Engineer
Back-end Developer & MLOps EngineerContributions:3 releases, 67 commits, 5 PRs in 6 months
Contributions summary:Justas's commits focus on modifying the core Python script for running the ProteinMPNN model. These changes include updating flags, adding options for calculating sequence-independent probabilities, and integrating additional CA-only models. The user also added a submission script for a bias experiment, showing a focus on model usage and experiment execution. Furthermore, the user added a flag to compute unconditional probabilities and introduced a random seed configuration, suggesting a contribution to model usability and experimentation.
deep-learningpytorch
dauparas/tensorflow_examples
Feb 2019 - Sep 2019
Contributions:76 commits, 80 pushes, 1 branch in 7 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.