My research focuses on AI in healthcare, with two main directions. First, I develop trustworthy and interpretable AI systems using explainable (XAI) models, reasoning frameworks, and visual grounding techniques to enable transparent and reliable clinical decision-making. Second, I advance human-centered AI by integrating human gaze data, medical images, and clinical reports to align model reasoning with clinicians’ workflows and improve diagnostic accuracy.

🌟 Key Skills:

• Programming: Python, Bash/Shell, SQL
• Machine Learning & Deep Learning: PyTorch, scikit-learn, NumPy, Pandas, Matplotlib, Seaborn
• Methods & Topics: Vision–Language Models (VLMs), XAI methods, Multimodal Fusion, Reasoning, Grounding
• Computer Vision: OpenCV, torchvision, Albumentations, Detectron (detection, segmentation, tracking), openslide
• LLMs & Agentic AI: Hugging Face Transformers, LlamaIndex, LangChain, RAG, PEFT (LoRA/QLoRA), vLLM, prompt engineering and AI reasoning

• CIG – Contrastive Integrated Gradients (Accepted at IEEE/CVF Winter Conference on Applications of Computer Vision (WACV), 2026)
  Official implementation of Contrastive Integrated Gradients: A Feature Attribution-Based Method for Explaining Whole Slide Image Classification (WACV 2026).
  📄 Paper: https://arxiv.org/abs/2511.08464
  🔗 Code: https://github.com/maianhpuco/CIG

• DIMVImputation
  A Python package for missing data imputation with support for multivariate and incomplete data scenarios.
  🔗 https://github.com/maianhpuco/DIMVImputation

• DualProtoSeg
  Prototype-based weakly supervised segmentation framework for histopathology images.
  📄 Preprint: https://arxiv.org/pdf/2512.10314v1 (under submission)
  🔗 Code: https://github.com/maianhpuco/DualProtoSeg

• Data & ML Blog
  Personal technical blog covering data-centric AI topics such as semi-supervised learning, domain adaptation, and practical ML workflows.
  🌐 https://maianh-learning.com/blogs/domain_adaptation/

• VAE Tutorials
  🔗 https://github.com/maianhpuco/VAE
  A hands-on tutorial on Variational Autoencoders (VAEs), including implementation details, comparisons with standard Autoencoders, and real-world applications such as topic modeling.

• Domain Adaptation Tutorials
  🔗 https://github.com/maianhpuco/learning-domain-adaptation
  A practical guide on handling domain shift using sample re-weighting techniques for robust model generalization.

🤝 Let's Collaborate:

• I'm always open to new collaborations and exciting projects. Feel free to reach out to me at maianhuel@gmail.com or connect on LinkedIn.

📫 Contact Me:

• 📧 Email: Your Email
• 🌐 Website: maianh-leaning.com

📊 GitHub Stats:

Thanks for visiting my profile! Let's create something amazing together. 🚀