I build systems for high-leverage problems.

Exploring the intersection of systems, tooling, and photonics.
Not a portfolio. This is a builder's lab.

Active Context

Most of what lives here is evolving, experimental, or being refined in public.

Parkinson’s Law: Work expands to the time alloted to it.

Selected Work

Context

Simulating and analyzing electromagnetic wave propagation in photonic structures efficiently.

Tension

Traditional photonics solvers are either slow or hard to scale across platforms and materials.

Decision

Built a high-performance Python Finite-Difference Time-Domain (FDTD) solver with GPU acceleration and flexible material models.

Outcome

Accurate, extensible waveguide photonics simulations that run fast on both CPU and GPU, making advanced optical modeling accessible.

Context

Document extraction is often brittle or prohibitively expensive at scale.

Tension

Generic OCR lacks structure; specialized models lack speed.

Decision

Built an opinionated region analyzer in Python focusing on layout topology first.

Outcome

High-fidelity extraction for complex docs without the ML overhead.

Thinking

If something here resonates,email me.

[email protected]