Hey, I’m Justin - an MS Robotics student (with a Bachelors in CS) at the University of Michigan.
I’ve worked on everything from autonomous VTOL drones to reinforcement learning agents with modern legged robots.
I enjoy tackling tough problems, especially if they involve robots or neural networks - anything remotely sci-fi is my jam.
University of Michigan | Ann Arbor, MI
M.S in Robotics (Class of 2026)
B.S.E in Computer Science and Engineering (Class of 2024)
Email:
lujust@umich.edu
Agile Robot Control and Design (ARCAD) Lab
Researcher (2025 - Present)
- Currently experimenting with novel RL control systems on the Unitree G1 humanoid and Unitree Go2 quadruped.
- Also developing experimental variants of Constrained-PPO algorithms based on papers that I find really cool and sick.
APR Beauty
Software Engineering Intern (June 2024 - Aug 2024)
- Integrated Microsoft Business Central ERP to automate inventory stock updates and enhance supply chain operations.
- Developed a data pipeline using Apache Spark and SQL to track supply chain metrics for major clients (Revlon and Estée Lauder).
Michigan Robotics
Robotics Intern (Feb 2023 - Aug 2023)
- Contributed to Dr. Wami Ogunbi’s research on stair climbing algorithms for Agility Robotics’ Cassie & Digit.
- Experimented with Reinforcement Learning in combination with traditional control-theory robot control methods.
The Globe and Mail
Software Developer Intern (May 2022 - Aug 2022)
- Worked on Sophi.io’s PostgreSQL queries and optimizations for content recommendation engines.
- Improved backend data workflows and API response times.
Miracotech
Software Developer Intern (June 2021 - Aug 2021)
- Built a deep learning-powered social media analytics platform using PyTorch and BeautifulSoup.
- Assisted in migrating systems to Google Workspace/Microsoft 365, website setup, and automating workflows.
President, Michigan Vertical Flight Technology (MVFT) | 2020 - 2024
- Led a 50+ member team to consecutive victories in the United States VFS Design-Build-Vertical Flight Competition.
- Developed flight software for a 2-meter wingspan heavy-lift VTOL aircraft for autonomous payload delivery.
- Secured $60,000 in funding through career partnerships with Ford, Blue Origin and Airspace Experience Technologies.
- Kineval is a Javascript / ThreeJS implementation framework for web-based robot kinematics, dynamics, and autonomous control.
- Kineval is compatible with ROSBridge and, as such, can be used to control actual robots through a lightweight interface.
- Robots tested on: MR2, Fetch, Spot, Cassie (WIP)
- We implement a real-time SLAM pipeline (particle filter + LIDAR) on an MBot, letting it map and localize while roaming around.
- Integrates A* path planning to explore frontiers, find exits, and return home - all autonomously.
- Implements solid motion control (PID-based wheel velocity + sensor fusion) to handle tight turns and avoid sensor drift.
- We present a 5-DOF Trossen Robotics arm paired with an Intel RealSense L515 for real-time block detection and for shooting hoops.
- Uses optimized closed-form forward and inverse kinematics solvers to stack and sort blocks (with surprisingly good accuracy)
- We also present a mini-basketball launcher for automated free-throws like Lebron (but not Shaq).
