Seattle Solvers FTC #23511
Software Team Member
Built a differential swerve drivetrain and custom path follower in Java (SolversLib, JUnit 5), plus a browser simulator that matches the Java on 166,400 random inputs and generates autonomous code from planned paths.
Cut mean final position error 6.8x (18.8 to 2.8 in) and mean autonomous time 16% (5.60 s to 4.71 s) over 42 runs.
Owned path planning in a three-way split of the autonomy stack with vision and routines; every change passed AI review, then programming-lead review.
Team code contributions and experiments across the Offseason, Decode, BioBuzz and Crumates programs, plus smaller tools.