Physics problems,
computational answers.
My work connects cosmology and gravitational-wave astronomy with statistical computation, scientific machine learning, and carefully validated software.
Surrogate inference for expensive likelihoods
Scientific ML & Bayesian inference
Surrogate posterior methods for expensive likelihoods, including GPry and its applications to cosmology and LISA.
02 LISA cosmology and primordial spectra
Early-universe gravitational waves
Using LISA to reconstruct primordial curvature perturbations and connect stochastic backgrounds to early-universe dynamics.
03 Non-Gaussian statistics and fast PTA inference
Pulsar Timing Arrays
What correlated nanohertz data can reveal about gravitational-wave source populations, plus accelerated and interoperable inference.
04 Typed graph planning and execution
Reliable AI systems
Typed planning, deterministic validation, local repair, and reusable execution with Graphsmith.
