I'm Peter
PhD Candidate @ The George Washington University
About Me
“All models are wrong, but some are useful.” — George E. P. Box
My work sits at the intersection of data science, machine learning, and large-scale simulation, with a strong emphasis on modeling complex, dynamic environments.
I am a PhD researcher with hands-on experience developing predictive, statistical, and learning-based models for high-dimensional spatiotemporal data. My work spans connected and automated vehicles, real-time sensing, deep reinforcement learning, computer vision, and statistical calibration, applied to large-scale data from sensing and connected systems.
Excited to connect and exchange ideas.
Experience
Developing AI-driven perception & planning systems and building advanced simulation pipelines for AVs
Machine Learning Engineer
The George Washington University
Washington, DC United States • September 2021 - Present
Experienced collaborating with federal research labs — FHWA Turner-Fairbank Research Center — to integrate advanced AI and AV algorithms into intelligent transportation and automation platforms.
Featured Projects
Jan 2024 – Aug 2026
NGM4AVs (Next Generation Modeling for AVs)
Developing motion and behavioral models for AVs in mixed traffic, integrating physics-based and learning-based methods including RNNs and deep reinforcement learning.
Aug 2023 – Aug 2025
AVA (Automated Vehicles for All)
Collaborative AV project with UIUC, Texas A&M, and UC Davis — full autonomous driving stack covering perception, sensor fusion, HD map generalization, and motion planning.
Aug 2022 – Aug 2024
TGSIM (Third Generation Simulation Data)
Large-scale naturalistic trajectory dataset collection using aerial and roadside cameras, with multi-object tracking, BEV transformation, and Kalman filtering pipeline.
Publications
Research contributions in AV perception, motion prediction, trajectory generation, and high-fidelity simulation
Diffusion Process-Based Model for Network Trajectory Propagation
IEEE Transactions on Intelligent Transportation Systems
January 2026
View Publication