Articles

Papers, scientific notebooks, and research visuals — biology, biophysics, and engineering.

• Predicting LNP Encapsulation Efficiency from Formulation Parameters — 2026-04-09

  An open dataset, an XGBoost baseline, and a physics-informed neural network for predicting lipid nanoparticle encapsulation efficiency from formulation parameters — with an honest look at why cross-paper generalization is hard.

• Dense vs. Mixture-of-Experts: Benchmarking Local LLMs on Intel Arc Battlemage — 2026-07-23

  Why a mid-range Intel Arc B70 can run a 120B-parameter Mixture-of-Experts model faster than a dense model a quarter of its size — and what that says about memory bandwidth as the real bottleneck in local AI.