Stan DeLaurentiis

Applied AI · Plasma Physics · New York

Now building in AI and plasma physics — mostly pointing transformers at turbulent fluids and seeing what they get wrong. Before this it was fluid simulations of binary black hole systems. Based in New York.

things i've built

sodium lamp — a live WebGPU model of a hydrogen/oxygen sodium-emission cell: rate-solved non-LTE populations, trapped 589 nm radiation, and a fully audited energy ledger, running in your browser — light the burner or read the source
vibefab — a conversational virtual fab: describe a chip structure in plain english, a GPU plasma-etch solver fabricates it, and you orbit the 3D result step by step — talk to the live fab or browse the runs
Quorum — adversarial peer review for research papers, with reviewers that write and execute Python to verify or falsify the numerical claims
master-annotator — drop in an annotated PDF, get a pull request on your paper's repo with the reviewer marks applied as LaTeX edits

latest writing →

May 2026

What a foundation model breaks on plasma

lab notebook →

Jul 2026	Emulating TCV edge turbulence: an architecture notebook
Jul 2026	Anchored rollouts: conditioning the emulator on sparse diagnostics

news

Mar 2025 Hertz Fellowship Finalist — top 5% of U.S. STEM PhD applicants
Feb 2025 First-author paper published in The Astrophysical Journal — relativistic binary precession
Nov 2024 Completed MPhil in Applied Mathematics at the University of Cambridge
Jul 2023 First-author paper published in Monthly Notices of the Royal Astronomical Society — gaseous dynamical friction in AGN discs
May 2023 Graduated Columbia College — B.A. Astrophysics with Honors, Magna Cum Laude; Phi Beta Kappa
Mar 2023 Awarded NSF Graduate Research Fellowship
Jan 2023 Named Gates Cambridge Scholar — press: Columbia Spectator · URF · Columbia College · Physics Dept
Oct 2018 Semi-finalist at the International Chinese Bridge (汉语桥) Mandarin Competition — U.S. team representative, top 20 of 110+ countries · press