Yanuo Zhou

Hi! I’m Yanuo (Arnold) Zhou (周亚诺), a Researcher and ML Engineer building multi-agent systems and AI applications for health and biomedicine. I hold an MSc in Precision Health and Medicine from the National University of Singapore.

I am actively seeking PhD opportunities to advance AI-driven innovations in biomedicine and precision health.

Research Interests

• AI Agents & LLM Applications: Multi-agent LLM pipelines for regulatory writing and literature review, with deterministic checks against fabrication.
• AI for Health (AI4Health): Applying machine learning to healthcare challenges, from digital phenotyping to drug discovery and precision medicine.
• Digital Phenotyping & Wearable Sensing: ML models for continuous health monitoring, circadian rhythm analysis, and behavioral pattern recognition from wearable and smartphone sensors.
• Bioinformatics & Multi-Omics: Multi-evidence integration of transcriptomics and metabolomics for biosynthetic pathway gene discovery; RNA-seq biomarker identification.
• AI-Driven Drug Discovery (AIDD): Multi-architecture GNN benchmarking for drug–target link prediction on biomedical knowledge graphs.
• Biomedical Signal Processing: Physiological signal processing (ECG) and wearable sensing.

Recent Work

• Cui (formerly LitScribe) (2026-present): evidence-grounded literature review and research-judgment tool. LitScribe (Jan–Jun 2026) was a literature-review pipeline over six academic sources with reviewer–synthesizer critique-and-revise and LLM-checked citation support; Cui (Jun 2026–present) redesigns it around user-written claims, immutable LLM challenge rounds and human adjudication.
• PrimeKG GNN Drug Discovery (2025–2026): drug–target link prediction benchmark of five GNNs (GAT, RGCN, GIN, GraphSAGE, GCN) and an MLP baseline on PrimeKG. V2.0 fixed split-level leakage (71.5% of test edges had their reverse in training) and made the evaluation protocol explicit (hard negatives, 51-way ranking).
• Multi-Omics AE-GNN (2025–2026): 5-stage pipeline for ranking candidate biosynthetic-pathway genes from paired transcriptome and metabolome data. Dual autoencoders (545M params) + GAT with multi-evidence ranking across 132K genes × 7K metabolites.
• Passive Sensing vs. Personality Questionnaires (2025–2026): head-to-head comparison on three public university datasets (N = 1,559) for mental-health prediction; personality beat sensing in all 11 primary comparisons. First- and co-corresponding-author manuscript under review at IEEE JBHI.

Key Technical Skills

• AI/ML: PyTorch • PyTorch Geometric • scikit-learn • GNN (GAT, RGCN, GIN) • VAE • Transformers
• AI Agents: LangGraph • DeepAgents • sentence-transformers • asyncio • Qwen/Claude/DeepSeek
• Trustworthy AI: LLM evaluation and audit design • Upstream-first pipeline auditing • Deterministic release gates
• Bioinformatics: RNA-seq • Knowledge Graphs (PrimeKG) • Multi-Omics Integration • KEGG Enrichment
• Signal Processing: Wearable Sensors • ECG • Real-time Waveform Display
• Reproducibility: Git • Docker • YAML Config Pipelines • HPC • Audit-ready Workflows

Selected Awards

• 🏆 PRECISE Scholarship, National University of Singapore (2023)
• 🥉 3rd Prize, 10th National Student Optoelectronic Design Competition (2022)
• 🥈 2nd Prize, 8th National Student Optoelectronic Design Competition (2020)
• 🥈 Silver Award, 12th “Challenge Cup” Jiangxi Student Entrepreneurship Competition (2020)

Get in Touch

Feel free to reach out via email at yanuo.zhou@outlook.com or connect with me on GitHub @arnold117.