Hi, I’m Yilin (Dylan) Wang—a Ph.D. candidate in the Lyles School of Civil & Construction Engineering at Purdue University, advised by Dr. Yiheng Feng.

About Me

With a multidisciplinary background spanning civil engineering (B.S., Tongji University), smart city and deep learning (M.S., Carnegie Mellon University), and intelligent transportation systems (Ph.D., Purdue University), I bring a unique cross-domain perspective to solving complex problems at the intersection of infrastructure, AI, and autonomy. My research lies at the intersection of cooperative perception, Physical AI, and autonomous driving, where I contribute to building intelligent systems that bridge academic research and industry deployment.

I completed a Simulation Development Engineer internship at Inchor (SaferDrive AI) (Dec 2025 – Jun 2026) and continue as a core developer of TeraSim, a city-scale agentic world model platform for autonomous vehicle safety validation. I am actively seeking Research Scientist / Applied Scientist / ML Engineer positions in autonomous driving and intelligent transportation systems, where I can leverage my interdisciplinary expertise to advance the frontiers of autonomous systems, cooperative perception, and AI-driven transportation.

You can find my CV here: Yilin Wang’s Resume

Research Interests

My research focuses on intelligent transportation systems and connected & automated vehicles (CAVs), with expertise in:

• LLM/VLM Agents & World Models for AV and Embodied AI: Trajectory-planning foundation models for behavior simulation (sim agents), world-model construction, and safety-critical scenario generation with closed-loop validation
• Cooperative Perception: V2X-based sensor fusion, infrastructure-vehicle cooperative perception systems, LiDAR and camera data processing
• Robotaxi & AV Fleet Operations: Routing, dispatching, and repositioning at network scale via MILP and learning-based combinatorial optimization, with fleets doubling as mobile sensing platforms
• Physics-informed Machine Learning: Integrating domain knowledge (e.g., car-following models) into deep learning for trajectory prediction
• Agentic System Implementation & Deployment: Multi-agent architectures, tool orchestration (MCP), and agent evaluation for production workflows, from smart-city and robotaxi operations to automated research pipelines

I integrate multi-modal data (e.g., UAV videos, LiDAR, V2X) with machine learning to enhance mobility, safety, and system optimality. I have led projects funded by agencies including FHWA, Michigan DOT, and CCAT, and am passionate about turning rigorous research into deployable systems that improve real-world transportation networks.