Technical notes, project writeups, and research-facing essays adapted from my ongoing work.
A technical blog adaptation of our IAQVEC 2026 accepted paper on using an LLM as an auditable workflow layer over lightweight urban physics models for PET, cooling demand, and design intervention analysis.
A technical blog adaptation of our AI4X manuscript on why structured-grid electromagnetics should be implemented as convolutions or tensor slicing instead of graph message passing, without giving up exactness.
A technical blog adaptation of my manuscript on why input enrichment worked better than cross-attention or output-side constraints for Newtonian discovery in long-context Transformers.