Alexey Radul

Papers

Total Citations

H-Index

About

Alexey Radul is a researcher specializing in probabilistic programming, Bayesian inference, and intelligent systems. His most influential work centers on making probabilistic programming languages more flexible and powerful — most notably through his development of **inference metaprogramming**, introduced in his 2018 paper "Probabilistic Programming with Programmable Inference" (collectively accumulating nearly 50 citations). This work challenged the prevailing paradigm of rigid, black-box inference algorithms embedded in language runtimes, instead empowering programmers to customize and compose inference strategies through new language constructs, earning it recognition as a foundational contribution to the field. Radul has also advanced the application of probabilistic programs to real-world reasoning tasks, including goal inference for autonomous agents — combining randomized path planning with probabilistic models to interpret the intentions of people, cars, and robots from partial observations. His 2015 work on Gaussian Process Memoization further demonstrated his commitment to making sophisticated statistical tools more accessible to practitioners across machine learning, robotics, and scientific computing. Taken together, Radul's research reflects a consistent vision: building expressive, programmable tools that bring the full power of probabilistic reasoning to complex, real-world problems.