> portfolio
Shenghua (Simon) Jin
Robotics @ CMU SCS · Class of 2029
Robotics student at Carnegie Mellon's
School of Computer Science, building
navigation, perception, and systems
software for real robots.
Currently on CMU RoboBuggy and Moon
Miners. Previously: ROS 2 navigation
and target detection at Ascension
Robotics, LiDAR–camera fusion research
at SFU, and a GPU compiler in OCaml.
scroll▾

// projects

Projects

Every face of the dodecahedron, browsable as a plain list.

// experience

Experience & Education

CMU RoboBuggy — Member

Sept 2026 – Present

Member of Carnegie Mellon’s autonomous buggy team, which builds self-driving buggies that race in CMU’s annual Sweepstakes.

CMU Moon Miners — Member

Sept 2026 – Present

Member of Carnegie Mellon’s Moon Miners, a student robotics team that builds lunar excavation and construction robots for NASA’s Lunabotics competition.

Ascension Robotics — Software Developer

Sept 2025 – June 2026

Wrote a TurretDecouple C++ ROS 2 component — a heading-locked tf2 frame plus remapped velocity commands — so Nav2 can plan for a holonomic base carrying a spinning LiDAR turret, then migrated the local controller from DWB to MPPI (omni model) to restore lateral motion. Raised the 8-class detector’s mAP@0.5:0.95 from 0.66 to ~0.98 by rebuilding the dataset (~7K → ~14K images), cut manual labeling ~75% with ONNX pre-annotation, cut camera-intrinsic RMS reprojection error ~80%, and moved bringup to a systemd user service — all on a Jetson Orin Nano.

LiDAR–Camera Sensor Fusion — Undergraduate Researcher, SFU

Jan – Apr 2026

Owned integration of a ZED stereo camera and a Robosense LiDAR on a 5-person team’s Jetson Orin Nano / ROS 2 Humble rig; calibrated LiDAR–camera extrinsics to under ~3 cm translation and generally <15 px reprojection error by seeding the 6-DoF optimizer with CAD measurements and ROI constraints, and designed the two-tier 3D-printed sensor mount.

Artiverse — Technical Data Annotator

Oct 2025 – Feb 2026

Automated URDF generation for 800+ .glb models with multi-threaded Python/NumPy scripts for physics simulation, and fixed legacy conversion scripts that silently corrupted output by isolating coordinate-system and baked-transform bugs.

Team Parabellum — Team Captain (FTC)

Sept 2022 – Jul 2025

Directed the mechanical design of the competition robot — custom chassis, capstan-drive speed reducer, and a telescoping box-tube linear extension with a differential wrist — carrying the team to 2nd Place Innovate at the European Premier Event; programmed Java autonomous routines with a custom PIDF controller and set the team’s Git/GitHub standards.

Education

Carnegie Mellon University

B.S. Robotics, School of Computer Science

Aug 2026 – May 2029 (expected)

Transferred from SFU · Robotic Autonomy (SLAM, motion planning, ROS 2, control theory), Linear Algebra

Education

Simon Fraser University

Computing Science (transferred to CMU)

2025 – 2026

CGPA 4.04/4.33

// skills

Skills

Robotics

ROS 2 (Nav2, MPPI, tf2, slam_toolbox)Camera/LiDAR calibrationMotion planningJetsonLinuxsystemd

Languages

C++PythonOCamlRustJavaJavaScript/TypeScriptSQL

Perception, ML & GPU

OpenCVYOLOv12ONNX RuntimeCUDA/NVRTCllama.cppNumPyscikit-learn

Hardware & Tooling

Fusion 360KiCadANSYS HFSScolcon/CMakeGitGitHub ActionsDocker

Web

ReactNext.jsTauriNode.js

// recognition

Awards