Applied Scientist
Treverse LLC
Applied Scientist / Xuhui Zhan
Research and production, in one loop.
Applied scientist across recommendation, multimodal learning, computer vision, language systems, and ML infrastructure, from research questions through deployed systems.
Practice topology
Six projects moving between questions, models, and systems.
Method / 05
Each stage opens the project that best demonstrates it.
01 / 05 Open project
Frame the question.
Ancient Mortars structures roughly 10 million particle images for comparison.
02 / 05 Open project
Fit the representation.
Inverse-LLaVA maps text into visual space inside the language model.
03 / 05 Open project
Make the comparison legible.
Negotiation transcript coding reports a cost drop from more than $5,000 to $3.
04 / 05 Open project
Publish a reviewable release.
FORGE defines six contracts across distinct ML runtimes.
05 / 05 Open project
Return field evidence.
Treverse joins controlled exposure with reproducible learning and serving.
Project index / 04
Selected projects.
Recommendation systems, multimodal learning, language systems, and computer vision.
Treverse Recsys: Temporal Recommendations from Learning to Serving
An end-to-end recommendation system combining inductive temporal graph retrieval, reproducible model releases, in-memory decision serving, and statistically controlled experiments.
Open system case study ↗ 
02 
03 
04 
See all projects Multimodal AI / Vision-Language Models
Inverse-LLaVA: Rethinking Multimodal Alignment via Text-to-Vision Mapping
03 Natural Language Processing
AI for Negotiation
04 Computer Vision
Ancient Mortars Classification
Experience / 03
Recent roles.
Lead AI Data Science Researcher
AI Negotiation Lab @ Vanderbilt University
Graduate Research Assistant
Network and Data Science Lab @ Vanderbilt University