The next chapterUniversity of Washington
Master of Science in Electrical and Computer Engineering
Signals → Perception → Intelligence
A Personal Research AlbumVol. 01 / 2026
AI Researcher & Engineer
Chengtai Du.
From Signals
to Intelligence.
MSECE Student · University of Washington
Hi, I’m Chengtai Du, an MSECE student at the University of Washington. My interests and work span computer vision and reinforcement learning, with a focus on diffusion models and world models. I also love music. I’ve played guitar for four years and enjoy taking it on stage.
Current research interests
- Computer Vision
- Reinforcement Learning
- Diffusion Models
- World Models
CD—001Stereo / 33⅓ RPM
++
Scroll to explore the recordSeattle, Washington
02EducationThe foundations
Education.
An MSECE student at the University of Washington, building on a background in Information Engineering from Xi’an Jiaotong University and experience in visual perception, 3D reconstruction, and robotics.
The next chapter
Where it beganXi’an Jiaotong University
Bachelor of Engineering in Information Engineering
92.49/100Major GPA
Top 3%Among 203 students
Academic honor
A milestone along the way.National Scholarship (China)
03InternshipsFrom theory to practice
Internships.
From synthetic scenes to physical robots. Learning by building systems that work.
01
Hikvision
Jun — Sep 2026Perception and Reconstruction Algorithm Engineer Intern
Built a Blender-based synthetic data generation pipeline and a C++ structured-light stereo reconstruction module, validating reconstructed depth against NVIDIA Isaac Sim ground truth.
02
Coleap.AI
Mar — Jun 2026Perception and Reconstruction Algorithm Engineer Intern
Developed a SAM 3D-based metric-scale reconstruction pipeline using RGB-D data and geometric alignment, and deployed it on a wheeled robot for real-world demonstrations.
Explore Open-Source Code (opens in a new tab)03
HSBC Software Development
Aug — Nov 2025Emerging Talent Program Intern
Worked on Spark-based data pipelines, automated testing, and CI/CD workflows, delivering 20+ data tickets and developing 30+ unit tests with over 90% coverage.
Perception. Reconstruction. Engineering.
04ProjectsSelected work
Projects.
Selected work in visual perception, spatiotemporal learning, and robotics.
Spatiotemporal learning01
Event-Based Object Detection
Institute of Artificial Intelligence and Robotics, Xi’an Jiaotong University
Jun 2024 — Dec 2025Developed spatiotemporal learning modules, including C2f-LSTM and channel-wise learnable LIAF neurons, for event-based object detection.
48.1%mAP · Gen1
48.5%mAP · 1 Mpx
Vision meets robotics02
Industrial Measurement
Xi’an Jiaotong University Robotics Club
Jan — May 2024Built an RGB-D-assisted component segmentation and measurement system using YOLOv5-seg, contributing to a National First Prize in the China Robot Competition and RoboCup Robot World Cup China.
90%+Classification accuracy
100%Region assignment¹
National First Prize
China Robot Competition and RoboCup Robot World Cup China
¹ Spatial region assignment accuracy in the evaluated setup.
05OutroThe human side
A different kind of frequency
Beyond the Lab.
Life is about more than the next experiment. Music gives me energy, too. I’ve been playing guitar for four years, and I love bringing my Martin LX1E travel guitar along to perform. You can call him “Mr. Yellow.”
MusicTechnology
On the road, on stage
Martin LX1E · Travel guitarMr. Yellow.
A little guitar. A lot of stories.