ML Research Scientist (probabilistic inference)
Core
Develop and evaluate probabilistic inference methods, specifically amortized inference, to translate theoretical insights into practical implementations for AI safety.
Role type
Research Scientist (Probabilistic Inference & AI Safety)
Builds
Theoretical proposals and high-quality implementations of probabilistic models
Domain
AI Safety, Probabilistic Graphical Models, Deep Learning
Deliverable
production ML models | research
Required skills
Probabilistic inference, Bayesian inference, Amortized inference (Variational Inference, GFlowNets), Parameter/structure learning in probabilistic graphical models, Reinforcement learning, Optimal control, Python programming, PyTorch/TensorFlow
Preferred skills
PhD in relevant field, Exceptional abilities in deep learning research
Technologies
Python, PyTorch, TensorFlow
Responsibilities
Develop amortized inference methods for high-dimensional distributions, Design evaluation strategies for probabilistic inference methods, Collaborate with mathematicians on theory, Translate theoretical proposals into code, Analyze experimental results to steer research directions
Seniority
Senior, hands-on IC