solver surface
small cases before big claims
start plain enough that every boundary condition, flux, and update has nowhere to hide.
colin vargas
evidence-first aerospace software for numerical tools, solver experiments, and project records that make physical behavior easier to inspect.
computational aerospace / numerical methods / propulsion / technical records
solver specimens scroll readout
a calculation earns trust by becoming inspectable.
mesh pass
let refinement reveal the argument
tighten the same case and watch which numerical choices start to matter.
stability readout
show when the model stops behaving
push the step size until the readout stops pretending everything is fine.
signal engineering taste
the useful signal is the place where math, machine, and physical system all get a vote.
- numerics
- stability before performance theater
- propulsion
- trade spaces over hunches
- software
- systems that can be reopened later
- taste
- dense technical surfaces with a little blur at the edge
selected work full index
selected work reads best as evidence, not decoration.
adaptive cfd solver
scalar advection work kept small enough for boundary conditions, updates, and plots to stay inspectable.
mesh, state, flux, update active notebookscramjet analysis tool
propulsion sweeps shaped around assumptions, trends, and comparisons that can be argued with.
assumption, sweep, trend writeup pendingcompressible flow project
finite-volume outputs paired with enough provenance for the figure to mean something later.
finite-volume inspection archive pending solver trajectory from case to record
the record clarifies when each responsibility has an address.
make it inspectable
keep first cases small enough that a mistake has nowhere to hide.
case is small on purposeseparate the hard parts
let mesh, state, flux, and update paths carry their own responsibilities.
complexity has addressesleave a record
publish notes and figures when they explain the decision, not just the result.
evidence before polish working style repeatable records
the working rhythm is calculation, assumption, record.
build the calculation.
start with local tools, small verification cases, and source history that can be replayed.
name the assumptions.
write down boundary conditions, discretization choices, solver limits, and what changed between runs.
publish what sharpens it.
turn notes, plots, and logs into public pages only when they improve the inspection.
notes open log index
notes are not a content treadmill here. they are where project decisions, build scars, and numerical context get cleaned into something reusable.
about systems reasoning
aerospace engineering, numerical methods, propulsion, and a bias for tools that show their work instead of hiding behind a clean screenshot.