Photo by my lovely gf Liangyuting Zhang
Wenxuan Xu 许文轩
I am Wenxuan Xu, a Master's student in Computer Science at Dartmouth College. My current research focuses on MLsys and AI infrastructure, specifically LLM inference acceleration, where I contribute to the sglang project.
Previously, I conducted research in Human-Computer Interaction (VR/AR). While I now consider this a detour, it provided me with valuable research experience. I collaborated with Andrew Campbell, Hai-Ning Liang, Wolfgang Stuerzlinger, and Yuntao Wang, establishing a solid publication record.
I am actively seeking New Grad opportunities in AI infrastructure. If you are interested in my background, please feel free to contact me.
Education
Dartmouth College
Master of Science in Computer Science with Concentration in Digital Arts
Hanover, United States
Sep 2024 - Present
University of Liverpool (Xi'an Jiaotong-Liverpool University)
Bachelor of Science in Information and Computing Science
Suzhou, China
Sep 2020 - Jun 2024
• First Class Honours
News
2025 Dec 28
My first-author paper LENS has been released on arXiv! Check it out!
2025 May 01
I'm now starting to work as a Research Assistant at HealthX Lab, Dartmouth College with Andrew Campbell
2025 Jan 21
My two first-author papers have been accepted by IEEE VR 2025! Looking forward to seeing you in France!
Publications
LENS: LLM-Enabled Narrative Synthesis for Mental Health by Aligning Multimodal Sensing with Language Models
arXiv preprint
(2025)
Multimodal health sensing offers rich behavioral signals for assessing mental health, yet translating these numerical time-series measurements into natural language remains challenging. Current LLMs cannot natively ingest long-duration sensor streams, and paired sensor-text datasets are scarce. To address these challenges, we introduce LENS, a f...
Predicting Ray Pointer Landing Poses in VR Using Multimodal LSTM-Based Neural Networks
IEEE Conference on Virtual Reality and 3D User Interfaces
(2025)
Target selection is one of the most fundamental tasks in VR interaction systems. Prediction heuristics can provide users with a smoother interaction experience in this process. Our work aims to predict the ray landing pose for hand-based raycasting selection in Virtual Reality (VR) using a Long Short-Term Memory (LSTM)-based neural network with ...
Optimizing Moving Target Selection in VR by Integrating Proximity-Based Feedback Types and Modalities
IEEE Conference on Virtual Reality and 3D User Interfaces
(2025)
Proximity-based feedback provides users with real-time guidance as they approach an interaction goal. This type of feedback is par ticularly useful for tasks that require guidance during the interac tion process, such as selecting moving targets. This work explores proximity-based feedback types and modalities to improve the se lection of moving...
Exploring the Effects of Spatial Constraints and Curvature for 3D Piloting in Virtual Environments
IEEE International Symposium on Mixed and Augmented Reality (ISMAR)
(2024)
Piloting requires users to control and navigate the aircraft within a designated pathway, with a controller that utilizes two joysticks to control the aircraft. This task is representative of various daily and gaming scenarios, such as controlling the aircraft to capture the photo or navigating an object in a game from the start position to the ...
Warning
Problem: The current name of your GitHub Pages repository ("Solution: Please consider renaming the repository to "
http://".
However, if the current repository name is intended, you can ignore this message by removing "{% include widgets/debug_repo_name.html %}" in index.html.
Action required
Problem: The current root path of this site is "baseurl ("_config.yml.
Solution: Please set the
baseurl in _config.yml to "