Capabilities
Engineering Focus
The projects below share a practical structure: model the physical or runtime problem, implement the working path, and leave enough logs, metrics, or interfaces to inspect the result.
Technical Areas
Nonlinear solver design, RK4 integration, auxiliary residual construction, damping, convergence handling, and benchmark-minded implementation.
Seen in ballistic-solverBridging FAST-LIO odometry, C++ occupancy-grid mapping, parking planning, delay-aware control, runtime composition, and Ioniq CAN-facing tooling into one validated vehicle system.
Seen in autonomous parkingSerial transport, GNSS/RTK correction flow, buffered logging, watchdog-style recovery, and remote visibility under unstable conditions.
Seen in telemetry runtimeTurning raw operational data into segment-aware analysis, replay, metrics, and inspectable multi-pane workflows.
Seen in analysis GUICapability Matrix
| Capability | What Shows It | Main Evidence |
|---|---|---|
| Mathematical modeling | Turning physical behavior into a solvable computational model and residual formulation | ballistic-solver |
| API and deployability | C ABI, Python package, Unity/C# interoperability, explicit status outputs | ballistic-solver |
| ROS 2 / vehicle autonomy integration | FAST-LIO bridge, C++ pointcloud occupancy grid, planner integration, measured-delay follower, Ioniq CAN path, RViz control | Mapless Autonomous Parking |
| Operational robustness | NTRIP handling, serial retry, timeout detection, live monitoring, remote deployment workflow | Racing Telemetry Stack |
| Data analysis tooling | Track-core build, segment slicing, synchronized replay, pane-aware visual inspection | Racing Analyze GUI |
| Cross-domain synthesis | Collection → monitoring → analysis, or simulation → solver → packaging | Projects overview |
How I Tend To Work
- Problem first: I start from the engineering bottleneck before choosing the implementation path.
- Runtime matters: implementations should run, fail visibly, recover, and leave inspectable state afterward.
- Diagnostics matter: explicit statuses, replayable logs, plots, debug streams, and operational visibility are recurring themes.
- Interfaces matter: useful cores are packaged behind APIs, launch files, services, or analysis workflows.
Evidence Map
- Measured behavior: solver benchmark tables, regression scripts, and paper-level comparison metrics.
- Runtime integration: odometry alignment, native occupancy-grid updates, delay-aware control, command outputs, and real-vehicle parking validation.
- Operational tooling: logs, monitoring, replay, and analysis interfaces that make field behavior inspectable.