Extropic
Thermo ML Resident
$75K–$200K
San Francisco · On-site · Full-time · Posted Sep 8, 2026
- Experience
- No years specified; the residency is a minimum of 3 months.
- Visa sponsorship
- Offers are conditioned on authorization to receive controlled items under U.S. export control laws and regulations.
About the job
A resident researcher develops and evaluates probabilistic machine learning models and learning methods for Extropic’s thermodynamic computing hardware.
What you'll do
- Derive theory for probabilistic models and learning rules, including energy-based and diffusion models.
- Build experimentation infrastructure and implement, visualize, and evaluate model architectures, training algorithms, and benchmarks.
- Publish papers, contribute to open source, and share design insights with the hardware team.
What they ask for
- Experience with scientific Python and JAX or a similar deep learning framework (PyTorch, TensorFlow, or Keras).
- Strong foundations in probability and linear algebra.
- Projects or papers demonstrating hands-on applied machine learning and data science experience.
Nice to have
- Familiarity with deep learning theory and literature, including over-parameterization theory and scaling laws.
- Experience with energy-based models or diffusion models; graph neural networks; deep learning experimentation infrastructure; information geometry; computational Bayesian methods; or publications at top ML conferences.
Pay
$75K – $200K (currency and compensation basis not stated)
Tools
Python, JAX, PyTorch, TensorFlow, Keras, Slurm, Ray, Kubernetes
AI focus
Research and train probabilistic models, including energy-based and diffusion models, using a thermodynamic paradigm.
Model-written summary · the original posting is authoritative