Your second-worst review is probably all you need
We analyze nine years of ICLR reviews and find that Reviewer 2 is not the problem.

We analyze nine years of ICLR reviews and find that Reviewer 2 is not the problem.
We show that UCB is a quantile of the predictive distribution in disguise, elicitable by the pinball loss even though no scalar loss can elicit the moment formula directly. The …
We tour a chezmoi-templated dotfiles setup that bootstraps a fresh machine in one command — terminal, shell, prompt, multiplexer, multi-environment templating, and the developer …
We implement a fully functional Gaussian Process regression pipeline — Cholesky decomposition, posterior predictions, and gradient-based hyperparameter optimization — in pure …
Our paper "Empirical Gaussian Processes" was accepted to ICML 2026.
We study Empirical GPs, a principled framework for constructing flexible, data-driven Gaussian process priors. By estimating mean and covariance directly from a corpus of …
We present Ax, an open-source platform for adaptive experimentation built on BoTorch. Off the shelf, Ax achieves state-of-the-art performance across a wide range of synthetic and …
Our paper "Ax — A Platform for Adaptive Experimentation" was accepted to AutoML 2025 (ABCD Track).
Started as a Research Scientist at Meta on the Adaptive Experimentation team within Central Applied Science (CAS), based in New York City.
Submitted my PhD thesis, *Probabilistic Machine Learning in the Age of Deep Learning*, at the University of Sydney.