Open to Machine Learning, AI, and Data Analysis roles

Machine learning, AI, and data analysis for real-world visual data.

I’m Hongrui Zhu, a Biomedical Informatics M.S. graduate from Stony Brook University. I build and evaluate machine learning systems across computer vision, biomedical imaging, and data analysis—turning complex data into measurable, reproducible results.

View selected work Resume PDF GitHub LinkedIn

What I do

ML/AI problem solving with strong data analysis fundamentals.

I combine model development, data preparation, evaluation, and clear communication—skills that translate across machine learning, AI engineering, and data analysis roles.

I work on applied ML problems where the model has to survive real data: noisy images, sparse annotations, domain-specific constraints, and evaluation metrics that reflect downstream use.

My recent projects include CellSeg-UNICls for brain histopathology cell segmentation and classification, multi-class nuclei instance segmentation on PanNuke, and FishLEN for killifish length estimation. Across these projects, I built training pipelines, reproduced prior methods, integrated segmentation and representation models, engineered features, and evaluated results with task-specific metrics.

I’m currently seeking full-time opportunities in machine learning, AI, computer vision, and data analysis roles where rigorous experimentation, reliable implementation, and clear analytical thinking all matter.

Machine learning & computer visionSegmentation, classification, regression, feature extraction, and deep learning model development.
AI systems & applied toolingExperience with PyTorch/TensorFlow workflows plus Ollama, LangChain, RAG, and AI-agent tooling.
Data analysis & evaluationNumPy, Pandas, feature engineering, model evaluation, and metrics including Dice, AJI, PQ, macro-F1, weighted-F1, and regression error.

Education

Biomedical informatics, mathematics, and software engineering.

GPA 3.88 / 4.0

M.S. in Biomedical Informatics

Stony Brook University · Stony Brook, NY · Sep 2024 — May 2026 · Graduated

Thesis: CellSeg-UNICls: Decoupled Segmentation and Embedding-based Cell Classification for Brain Histopathology.

Double major

B.S. in Math & Applied Math and Software Engineering

Dalian Jiaotong University · Dalian, China · Sep 2019 — Jun 2024

Both graduation theses received University-Level Outstanding Thesis recognition.

Selected projects

Projects with concrete model outputs, not just topic labels.

Each project highlights the problem, technical approach, and measurable result behind the work.

Computational Pathology Graduation Project

CellSeg-UNICls

A decoupled framework for cell instance segmentation and embedding-based cell classification in brain histopathology images.

0.863Dice
0.555mPQ
0.933weighted-F1
0.747macro-F1
Cellpose-SAMUNI2MLPQuPathPyTorch
GitHub
Instance SegmentationPanNuke

MCV — Multi-Class Nuclei Instance Segmentation

Built a PanNuke nuclei segmentation pipeline with ResNet50 U-Net, binary mask prediction, HoVer-map regression, and PyTorch DistributedDataParallel for multi-GPU training.

0.859Dice
0.738AJI
0.716PQ
0.747AJI+
PyTorchU-NetResNet50DDPHoVer-map
GitHub
Scientific CVFishLEN

Fish Length Net

Developed a PyTorch-based system for killifish length estimation using fish segmentation masks, a U-Net with ResNet34 encoder, and a ResNet18 + MLP regression model.

87.56%baseline within 10% error
95.40%with metadata features
PyTorchU-NetResNet34ResNet18Regression
GitHub

Experience

Research experience across pathology, biology, and brain science.

A concise view of the lab context, technical contribution, and measurable impact behind each role.

Research InternshipJun 2025 — May 2026Department of Biomedical Informatics · Stony Brook University

Cell segmentation and classification for brain histopathology

  • Developed CellSeg-UNICls for cell instance segmentation and classification in brain histopathology images.
  • Built a two-stage pipeline combining Cellpose-SAM segmentation, UNI2 embeddings, and an MLP classifier.
  • Created QuPath/SAM-assisted annotations from sparse point labels to support limited-supervision training.
  • Achieved 0.863 Dice, 0.555 mPQ, 0.933 weighted-F1, and 0.747 macro-F1.
Research InternshipJan 2025 — Jun 2025Dormant Biology Lab · Stony Brook University

Computer vision system for fish length estimation

  • Developed FishLEN, a PyTorch-based computer vision system for killifish length estimation.
  • Used fish segmentation masks to support accurate length regression.
  • Improved predictions within 10% error from 87.56% to 95.40% using metadata features.
Research InternshipNov 2024 — Dec 2024Department of Biomedical Informatics · Stony Brook University

Reproduction and analysis of cell segmentation models

  • Reviewed and analyzed state-of-the-art cell segmentation and classification models.
  • Reproduced experiments from published work by running Hover-Net GitHub implementations.
  • Visualized extracted features and converted results into QuPath-compatible datasets.
Research InternshipOct 2023 — Dec 2023Center for Innovation in Brain Science · University of Arizona

TensorFlow migration and pipeline refactoring

  • Migrated a TensorFlow v1 codebase to TensorFlow v2, improving maintainability and stability.
  • Refactored training pipelines and removed approximately 20% redundant operations.
  • Reduced runtime and memory consumption by approximately 10%.

Skills

A practical toolkit for ML, AI, and data analysis workflows.

Programming

PythonJavaCSQLMATLAB

Machine Learning

PyTorchTensorFlowCNNResNetU-NetMLPSAM

Computer Vision & Analysis

Image SegmentationInstance SegmentationComputational PathologyFeature EngineeringModel Evaluation

Tools & AI Systems

GitGitHubQuPathMySQLSQL ServerOllamaLangChainRAGAI Agents

Contact

Let’s talk about ML, AI, data analysis, or computer vision roles.

I’m happy to connect about full-time opportunities, research collaborations, and applied machine learning projects.

Email me GitHub LinkedIn Resume