Yuxin Ding

Projects

About Me

Portrait of Yuxin Ding

Yuxin Ding

Machine Learning Engineer · Medical AI · Computer Vision

I build machine learning systems for turning real-world needs into practical, impactful products. At Carina Medical, I work across AutoBrachy’s pipeline, from multi-site CT/MR data processing and 3D segmentation to geometry-aware reconstruction, dose optimization, and evaluation supporting FDA 510(k) submissions.

Beyond work, I enjoy exploring new domains and data types. My projects include a local RAG chatbot for product manuals, an AI agent for adjusting radiotherapy plans through natural-language instructions, and earlier work involving traffic video analysis, speech and text processing, and epidemic modeling.

400+
clinical cases in the ML pipeline
5–6
institutions' data harmonized
3+
years building medical AI
AWS Certified Machine Learning Engineer - Associate badge AWS
Certified ML Engineer – Associate (2026)

Experience & Education

Sep 2023 – Present Data Scientist Carina Medical LLC · Baltimore, MD / Remote
  • Maintain and extend AutoBrachy's ML data and training pipeline across 5–6 institutions and 400+ clinical cases, with data QA, ROI-definition reconciliation and detection of corrupted, duplicated or mismatched DICOM data.
  • Reproduced out-of-distribution segmentation failures and built simulation-based augmentation for them, raising tandem-and-ring processing success from 129/145 (89%) to 145/145 (100%).
  • Developed geometry-aware post-processing for vertebral separation, applicator reconstruction and merged-needle separation (trajectory prediction, Hungarian matching, spline reconstruction).
  • Implemented dose calculation and constrained dwell-time optimization, plus reproducible multi-site evaluation supporting FDA 510(k) documentation.
Sep 2022 – Dec 2023 M.S.E. in Applied Mathematics and Statistics Johns Hopkins University · Baltimore, MD
  • Teaching assistant for EN 553.636 Introduction to Data Science (Spring 2023).
Jun – Aug 2021 Software Engineering Intern VMware · Beijing, China
  • Built Python tooling to monitor and visualize algorithm outputs and long-running jobs, and benchmarked clustering metrics and bin-packing algorithms for algorithm selection.
Sep 2018 – Jun 2022 B.S. in Statistics, Minor in Data Science Fudan University · Shanghai, China

Skills

blue skills link to projects. Click one to see where I used it.

Machine Learning & Computer Vision

PyTorch scikit-learn OpenCV 3D segmentation Object detection & tracking Sequence models Medical image analysis Data augmentation Model evaluation

LLMs & RAG

RAG LLM agents LangGraph Hybrid retrieval Local LLMs (Qwen3, Ollama) FastAPI

Optimization & Statistics

CVXPY Numerical optimization Stochastic simulation NumPy SciPy pandas Hypothesis testing Non-inferiority analysis

Medical Imaging & Radiotherapy

DICOM CT / MR imaging SimpleITK pydicom RTSTRUCT / RTDOSE TG-43 dose calculation DVH analysis
Languages & tools Python SQL Java R Django Git Linux Docker

Contact