GCSP · Multidisciplinary
2026
Sim-only test-stand software: abort profiles encode risk tolerance (profile_b at 380 psi vs default 420 psi), not a single failure point.
HotFire-Ledger is my Multidisciplinary anchor: sim-only propulsion test-stand software spanning controls, human-factors process design, and information integrity.
Two abort profiles exist because conservative operators may want to cut power earlier than a vehicle's failure threshold; profile_b aborts at 380 psi vs default 420 psi. Role locks and a hash-chained ledger record who signed off on what.
What I worked on during 2026. Hover underlined terms for quick definitions.
HotFire-Ledger is a GCSP reflective capstone codifying Rocket Propulsion LaboratoryRPL hot-fire discipline in a sim-only repo. FSM logic, role locks, hash-chain verification, and reduction formulas are real (54 pytest tests). Physical propellant, sensors, and a live test stand are out of CI scope by design.
At Rocket Propulsion LaboratoryUSC liquid and hybrid propulsion test campaigns I write logging, checklist, and abort tooling for real hot-fire days. Software bugs there are not redeploy-and-fix events; missed over-pressure readings are hardware and safety failures.
BenchHost answered who can afford a benchmark. HotFire-Ledger answers who may sign off on which step and whether the log survives a dispute after the run.
Abort pressure is a judgment about risk tolerance, not a single physical failure point. profile_b cuts chamber pressure earlier than the default profile because a conservative safety_officerRole that can veto or override operator actions on test day may want margin a designer considers excessive (380 vs 420 psi in the sim ontology).
Checklist role locks prevent one person from arming and signing off alone. The hash-chained ledgerEach entry hashes previous entry + payload; tamper test names first broken index exists so a dispute after the run has a record both sides can verify, not a shared spreadsheet someone edited.
Abort limits and ICD thresholds both encode institutional risk posture. NASA internship interface work pushed me to StandardsLintMulti-agency ICD linter to catch cross-agency conflicts before a joint review.
Software bugs on a test stand are not redeploy-and-fix events. Role locks and hash chains are part of the experiment record. Multidisciplinary discovery requires controls, human factors, and information integrity in one auditable path.