I build simulations.
Then I prove them.
A one-person simulation lab. GPU physics from first principles, each validated to a number against an independent reference — then released, finished.
The work.
Each release is a simulation validated to a number against an independent reference, then shipped as an instrument you can run in your browser. The canvas above is live GPU physics, not a recording.
$ ay4la releases --count
2 shipped · all live · 1 in the kitchen
- 01
Brasa
synthetic wildfire thermal imagerySynthetic longwave-IR imagery of wildfire scenes — fires burning on real 1 m airborne-lidar ground, scene radiance modeled from first principles, then run through a physical sensor model. Training and test data for thermal fire detectors, built for the night-and-fog conditions where real footage runs out.
12 / 12physics stages that reproduce an independent reference — libRadtran, Rothermel, USGS 3DEP, FVS Sierra allometry, real FLAME 3 fire data. the weakest one (background litter texture, 1.54× rougher than real forest floor) is named on the release page rather than buried. - 02
Frostline
phase-change heat transfer · ISRUA phase-change heat-transfer engine for off-world resource work — melting and sublimating fronts moving through regolith and ice. Every moving boundary is gated against the exact Stefan solution, so what you get is validated physics, not a plausible-looking animation.
4e-4moving front vs the exact Stefan solution
A multiphysics engine coupling reaction kinetics with heat transfer — matter that changes state and composition as energy moves through it. Taking shape now; it will ship the way the others did, with a browser instrument and a number it's validated against.
The engine the others grew out of — GPU reaction-diffusion for cardiac electrophysiology, calibrated to published physiological bands. Not published as a release yet, but the solver lineage behind Brasa and Frostline starts here.
- radiative transfer
- Planck radiometry
- IR sensor model
- Stefan / phase change
- conjugate heat transfer
- reaction–diffusion
- GPU compute · WGSL
- reference-validated
How the lab
works.
Three habits, applied to every project — the releases and the commissions alike:
Validated, not asserted.
Every simulation here is checked against an independent reference — measured data, published laws, exact analytic solutions — before it ships, and the comparison goes with the work. Not "physics-based" as an adjective: validated to a number you can take into a design review.
First principles. No black box.
I model the radiometry and the sensor degradation myself, from first principles — no licensed engine I can't open, no generative layer inventing photons that were never there. Every stage is named and individually validated; every assumption is inspectable. You can audit the data you train on.
Talk to the engineer, not a funnel.
No demo funnel, no BDR gauntlet, no per-seat license. You work directly with the engineer building your project — the one who writes the physics and signs the validation report. Fixed-scope, fast iteration, direct access. More accountable than an enterprise platform contract.
Commissions.
The releases are free. If you want the lab pointed at your problem, I take on a small number of commissions when the physics is interesting. Most begin as one fixed-scope piece — you keep the tool, the code, and the validation report. Need something larger, or ongoing? That's a conversation I'm open to. Start by telling me the problem.
$ ay4la commissions --status
status: open
pricing: scoped to the brief
brief: elroy@ay4la.com
Synthetic thermal & IR data
Labeled synthetic IR/thermal imagery for training and testing detectors. Scene radiometry modeled from first principles, then degraded through a physical sensor model — every frame's radiance benchmarked against libRadtran before delivery. For detection that has to work at night, in fog, at 8–12µm, where real data is scarce.
- Perfectly labeled synthetic IR/thermal frames + label manifest
- Full sensor pipeline — optical PSF, NETD, FPN, ADC
- Scene radiometry: Planck emission, atmospheric path radiance
- Your sensor spec, your bands, your scenario
- Radiance benchmarked against libRadtran reference
- Plume/transport physics matched to published laws
- A validation report ships with the data — not just a claim
Validated multiphysics
Physics-correct heat-transfer and phase-change simulation for R&D that off-the-shelf tools handle poorly or slowly. Moving sublimation fronts, coupled thermal fields, energy/mass ledgers that actually close — each gated against an exact analytic solution before it's trusted.
- Conjugate heat transfer, conduction, radiation
- Phase change / moving-boundary (Stefan) problems
- GPU-accelerated fields — real-time where it matters
- Matched to analytic solutions + measured data
Computational biophysics
Reaction-diffusion modeling for biomedical R&D — cardiac electrophysiology and excitable-media problems, GPU-solved and calibrated to measured physiology. If it can be written as coupled fields and checked against a reference, it fits here.
- Cardiac electrophysiology (monodomain, BOCF)
- Source–sink capture, propagation, pacing
- GPU reaction-diffusion on structured grids
- Calibrated to published physiological bands
Not a clean fit? Say what you're after anyway — bigger builds and ongoing work are on the table.
Elroy Ayala.
corpus christi · texas
Principal of AY4LA LLC, and the only engineer on it. I build GPU physics simulations from first principles and validate each against an independent reference before it ships. One person owns the model, the validation harness, and the data pipeline — one line of accountability, no black box you inherit.
Rust on wgpu, GPU compute, custom solvers. Radiative transfer and IR sensor modeling, heat transfer and phase change, reaction–diffusion. Every result is benchmarked to a number — clear-sky radiance to libRadtran within 0.001 K, phase-change fronts to the exact Stefan solution — and the report ships with the work. The work section above is the record.
Let's build something.
A commission, a commercial license for a release, a bigger Brasa bundle, or just a physics problem worth talking about — write directly. You'll reach the person who does the physics.