News
- May 2026 NewOur paper Distilling Photon-Counting CT into Routine Chest CT through Clinically Validated Degradation Modeling received an early accept at MICCAI 2026.
Active projects
AutoMedBench
UC Santa Cruz × NVIDIA · 2026–present
Towards medical AutoResearch
- A workflow-aware benchmark that puts a base LLM in the driver's seat across the full automated research pipeline: planning, setup, validation, inference, submission.
- Five research categories — segmentation, enhancement, VQA, report, detection. 7 models; 5,400+ runs; two-container isolation.
- Uses LLM-as-judge scoring and curated skill sets to evaluate agentic ability across two scaffolding tiers — Lite (step-by-step guidance) and Standard (open-ended, minimal hints).
- Key finding: over 99% of error codes triggered are engineering-based; a single error code causes an average 48% score drop.
OpenVAE
UCSC × UCSF × Johns Hopkins × NVIDIA · 2025–present
Scaling latent backbones with worldwide data
- Pretrained 2D and 3D KL-VAE / VQ-VAE backbones for CT and MRI volumes.
- Trained on 400,000+ CTs collected from 145 hospitals worldwide; used as drop-in priors for downstream medical generative work.
Publications
- AutoMedBench: Towards Medical AutoResearch with Agentic AI Models.
- Distilling Photon-Counting CT into Routine Chest CT through Clinically Validated Degradation Modeling.
- See More, Change Less: Anatomy-Aware Diffusion for Contrast Enhancement.
- ShapeKit: Shape-Aware Postprocessing for Organ Segmentation.
- AI-Powered Translation Tool to Enable Contrast-Aware CT Synthesis.
- Towards Robust Out-of-Distribution Generalization via Bayesian Optimization.
Experience
Research Scholar — Johns Hopkins, CCVL
2025 – Mar 2026
Advised by Prof. Alan Yuille (Bloomberg Distinguished Professor)
- SMILE — anatomy-aware diffusion for CT contrast enhancement. ~50% FID gain, ~10% F1 lift on early tumor detection. CVPR 2026 (under review).
- ShapeKit — shape-aware postprocessing for organ segmentation. >10% average DSC improvement. MICCAI 2025.
Research Intern — Kermen Lab, UCPH
Oct 2024 – May 2025
Neuroscience Department · advised by Prof. Florence Kermen
- Designed ZATA-1.5B, a multi-attention zebrafish tracking model with transformer backbones for fine-grained behavior — ~80% sensitivity, ~90% specificity.
- Manuscript in preparation (PNAS).
Quantitative Research Intern — Shanghai Quantinv
Mar – Aug 2024
Investment & Decision Team · return offer
- Designed QuantNet-6B, an attention-based factor model with transformer backbones — 180% excess return at 20% maximum drawdown.
Algorithm Research Intern — Shanghai AI Lab
Oct 2022 – Oct 2023
Advised by Prof. Nanyang Ye, SJTU
- Parallelized Bayesian optimization for interpretable deep models — +50% accuracy at −90% runtime. First-author publication.