博士在读，具有统计学与计算金融背景。关注机器学习如何理解真实系统，并将研究问题转化为可复现的实验、模型与软件。
I'm a doctoral student with a background in statistics and computational finance. My work connects machine learning, physical modeling, and research software across sensing, semiconductor processes, healthcare, and microfluidics.
Research portfolio · Public repositories
| Area | Questions I work on |
|---|---|
| Intelligent sensing · 智能传感 | 气体传感、温度调制与长期漂移；研究表征中的有效信息及跨条件泛化。 Gas sensing, temperature modulation, drift, and representation learning under changing conditions. |
| Semiconductor processes · 半导体工艺 | 刻蚀、光刻与纳米压印中的物理建模、数值模拟和优化。 Physical modeling, numerical simulation, and optimization for etching, lithography, and nanoimprint processes. |
| Clinical machine learning · 临床机器学习 | 血液透析风险与纵向治疗数据；关注历史信息的增量价值和外部验证。 Dialysis risk modeling and longitudinal treatment data, with a focus on incremental predictive value and external validation. |
| Intelligent microfluidics · 智能微流控 | 液滴检测、图像分割、路径规划与实验设备集成。 Droplet detection, image segmentation, path planning, and integration with experimental instruments. |
Physical modeling · Monte Carlo · Scientific computing
面向氦离子光刻的多物理场蒙特卡洛研究框架，连接材料表征、离子输运、能量沉积与光刻粗糙度分析。
A research framework connecting material characterization, ion transport, energy deposition, and line-edge roughness analysis in helium ion lithography.
MCP · Remote research workflows · Python
面向远程 Windows 设备的自托管 MCP 网关，支持分级启用操作能力、文件访问与工作流核验。
A self-hosted MCP gateway for remote Windows machines, with tiered capabilities, file access, and workflow verification.
Research portfolio · Bilingual interface · HTML / CSS / JavaScript
汇集研究方向、学术论文与近期项目的双语主页，支持论文搜索、主题筛选与 BibTeX 引用。
A bilingual research portfolio with publication search, topic filters, project summaries, and BibTeX citations.
- Modeling & analysis — Python, NumPy, pandas, SciPy, scikit-learn, LightGBM, PyTorch.
- Scientific computing — Statistical modeling, numerical simulation, Monte Carlo methods, Gaussian processes, and Bayesian optimization.
- Computer vision — OpenCV, YOLO, SAM, segmentation, and morphological analysis.
- Research engineering — FastAPI, hardware SDK integration, experiment workflows, and AI-assisted development.
从具体问题出发，重视评估设计、可复现性与证据边界；让算法、物理知识和实验系统共同服务于研究目标。
I start with concrete questions, make evaluation assumptions explicit, and distinguish predictive evidence from causal claims. I aim to build research tools that can be inspected, reproduced, and used.
Stay curious. Follow the evidence.
保持好奇，尊重证据。