Resume

Summary

Zimu (Zim) Gong

M.S. CSE @ UM-Ann Arbor | ex Research Intern @ Waabi | Lightwheel's first intern

Second-year Computer Science and Engineering master's student at the University of Michigan-Ann Arbor. Interested in VLA models, flow matching, reinforcement learning, and simulation in autonomous driving and robotics. Hands-on experience in building data pipelines, designing and training machine learning models, conducting closed-loop evaluations, and exercising sim2real domain transfer.

Education

Master of Science (M.S.), Computer Science

2024 - 2026

University of Michigan-Ann Arbor, Ann Arbor, MI, US

Bachelor of Science (B.S.), Electrical and Computer Engineering

2020 - 2024

UM-SJTU Joint Institute, Shanghai Jiao Tong University, Shanghai, CN

Study Abroad, Electrical and Computer Engineering

2023

Cornell University, Ithaca, NY, US

GPA: 4.00 / 4.00

High School Diploma

2017 - 2020

上海市实验学校 (Shanghai Experimental School), Shanghai, CN

Languages

  • English — Native or bilingual proficiency
  • Chinese — Native or bilingual proficiency

Professional Experience

Research Intern

May 2025 - Aug 2025 · 3 mos

Waabi · San Francisco, California, United States

  • Working with A&A Behaviors Team.

Autonomous Division Member

Aug 2024 - May 2025 · 9 mos

MRacing · Ann Arbor, Michigan, United States

  • Working on Mapping and Localization. Designed and implemented EKF-SLAM algorithm on ROS for FSAE autocross event.

Graduate Student Instructor

Aug 2024 - Dec 2024 · 4 mos

University of Michigan · Ann Arbor, Michigan, United States

  • Graduate student instructor of SI650/EECS549 Information Retrieval FA24 with Professor Ceren Budak.

End-to-end Autonomous Driving Intern

May 2024 - Aug 2024 · 3 mos

Lightwheel

Returned to Lightwheel AI in summer 2024. Working closely with colleagues, we developed an in-house simulator that supports Unreal Engine and 3DGS hybrid rendering, which later served in the development of an end-to-end autonomous driving algorithm for an OEM.

  • Designed vehicle kinematic and dynamics models for the in-house simulator and developed PID and iLQR controllers for closed-loop evaluation.
  • Designed the navigation head that encodes raster map navigation information into embeddings that improve the planning robustness of the end-to-end algorithm (VAD).
  • Evaluated the effectiveness of hybrid-rendering data on models by training VAD using data collected from 3D Gaussian Splatting reconstructed scenarios from the nuScenes dataset in an in-house simulator.
  • Created a CI workflow that tags and builds releases of Python packages and Docker images for deployment.

Research Assistant

Jan 2024 - May 2024 · 4 mos

University of Michigan Transportation Research Institute · Ann Arbor, Michigan, United States

Developed a charging station recommendation system with Professor Zhen Hu and Professor Shan Bao at UMTRI.

  • Conducted feature extraction and significance analysis in 30 dimensions with LLM and SHAP on 2,000 Google Maps charging station review data.
  • Implemented a recommendation algorithm with a matrix factorized collaborative filtering model and utilized feature bi-encoding from LLM extraction to solve the cold start problem.
  • Designed a charging station search engine with more than 50,000 station data and built a web server with Flask.

Drive Sim Intern

Jun 2023 - Aug 2023 · 2 mos

Lightwheel

  • Designed an autonomous driving simulator with Unreal Engine as the renderer for both open-loop and closed-loop evaluation of end-to-end algorithms, using a ROS-like framework with proto and gRPC services for communication.
  • Constructed a data visualization system for custom autonomous vehicle simulation data (bounding boxes, maps, diagnostics) with the Python Foxglove framework.
  • Developed and maintained a Python library that contains sensor models and 3D transformation methods.