Case studies and live engineering tools built over 11+ years of independent work spanning CFD, FEA, optimization, automation, and full-stack software.
Enter any 1D or 2D mathematical expression and find its global minimum using differential evolution — the same optimizer class used in production engineering programs.
Input engine specs, boat speed, and water temperature to identify all viable heat exchanger configurations from a surrogate-modeled design space.
Parametric pressure drop predictions for custom heat exchanger geometries, derived from 125 OpenFOAM simulations and validated within 5% across held-out configurations.
Automated null surface design for GRIN and external glass lens inspection. Reduced an 8+ hour VP-level manual process to under 5 minutes.
Replaced a manually-tuned, geometry-dependent simulation with a generalized predictive digital twin accurate to within 50µm across arbitrary material stacks.
First 3D shaftline model for U.S. Navy aircraft carriers, developed in collaboration with NAVSEA. Revealed failure modes invisible to legacy 2D methods. Co-published at SNAME 2023.
Delivered load-dependent bearing stiffness values to customers for rotordynamic analysis — a capability that had never existed at the company before.
Used ANSYS thermal simulation to prove that a 1°C temperature variation across a heated tool was fully explained by thermocouple placement geometry — not a real non-uniformity defect.
Thermal cycling and multi-method shock simulations (SRS, modal superposition, transient) for weapon-mounted optical systems. Identified and corrected critical errors in a partner's simulation model.
Interactive differential evolution optimizer — enter any mathematical expression and find its global minimum. Runs a live FastAPI backend with SciPy.
Automated null surface design for GRIN and external glass lens inspection. Reduced an 8+ hour VP-level manual process to under 5 minutes via optimization scripting.
Surrogate model trained on 125 OpenFOAM simulations predicts pressure drop from geometry inputs. Validated within 5% on 25 held-out configurations.
Input engine specs and operating conditions — the tool identifies all viable heat exchanger configurations using integrated heat transfer and pressure drop models.
Custom application that generates complete CNC toolpath instructions for Navy stave bearings from a single set of parameters. Increased output 3–6× over manual production.
Full-stack ASP.NET + Angular customer portal with authentication, per-user usage controls, and automated 3D STEP file delivery — built entirely from scratch.
Three-phase U.S. Army-funded program: closed-loop molding optimizer, neural network surrogate reducing solve time from weeks to hours, and end-to-end manufacturing pipeline.
Designed and commercialized a complete 4"–15" dry-running seal product line from scratch — expanding a 2-SKU copied prototype into a differentiated, fully manufactured commercial family.
Custom ANSYS-designed seal handling 4× the commercial displacement spec. Prototype lip pressure matched simulation exactly. Still in active service 5+ years after installation.
Published Android app that uses a phone camera to match Philips Hue lights to on-screen colors in real time. 1,100+ downloads, $1,500 in revenue on the Play Store.
Independently pitched and secured two separate U.S. Army-funded programs — Supercharge and Watson — covering GRIN lens optimization and end-to-end manufacturing automation.
Co-developed the first 3D shaftline model for U.S. Navy aircraft carriers with a NAVSEA engineer. Results confirmed a suspected real-world failure mechanism and were published at SNAME 2023.
Restored a non-functional production system at a $25B+ manufacturer after the original developer and internal experts had failed. Machine was back online within days of a month-long outage.
Delivered working compressible flow simulations and full analytical solutions for a $2B+ manufacturer whose internal CFD team could not produce valid results.
Designed a targeted simulation to determine whether an experimentally observed phenomenon was physically real or an instrumentation artifact. Conclusively confirmed it was real.