Agentic AI · Computational Biology

I build AI agents that reason over single-cell data. They plan analyses, run tools, test hypotheses and show the evidence behind what they find.

Explore my research → Get in touch

Ph.D. student, Biomedical Engineering · Xia Lab · Advisor Prof. Zheng Xia

Fig. 1 · Human INS, bp 1–36B-DNA · 10.5 bp per turn · 0.34 nm rise

I design AI systems that read cells, form hypotheses, test them with real tools, and show the evidence behind every claim.

01 · Single-cell agents

Agents that read the tumor microenvironment.

At the Xia Lab I am building an autonomous, LLM-powered agent system for cancer single-cell analysis. It plans multi-step analyses, tests hypotheses iteratively and runs analysis tools on its own.

  • AnnotationA cell-type annotation agent that pairs biological knowledge retrieval with multi-step reasoning to characterize rare, transitional and malignant cells.
  • DiscoveryA reasoning agent for oncogenic gene discovery that infers cell–cell communication, functional gene perturbations and pathway shifts.

Xia Lab · UT Austin · 2026–present

Simulated single-cell embedding, for illustration

02 · Explainable reasoning

Every conclusion traces back to evidence.

An explainable-AI layer records each agent decision and tool call, then links it to verifiable biological evidence. A gene signature should arrive with its rationale.

  • TraceabilityDecision paths and tool executions mapped to the evidence that supports them.
  • BenchmarksReasoning fidelity, tool-use precision, self-reflection loops and hallucination rates, measured against expert-curated genomic datasets.

Xia Lab · UT Austin · 2026–present

Multi-agent system, schematic

03 · Earlier work: clinical AI

Clinical AI, tested against clinicians.

Before the Ph.D. I built and evaluated AI for reproductive medicine at the Human-Centered AI Lab, Pittsburgh Institute – Sichuan University.

87%clinician preference rate for ARTLLM in clinical reasoning and treatment planning
  • ARTLLMA decision-support system for infertility treatment, built with reinforcement learning.
  • EvaluationBlinded benchmarks of LLM clinical reasoning, and the “Alignment Paradox” between algorithmic and clinical performance.

Blastocyst: trophectoderm, inner cell mass, zona pellucida

5peer-reviewed papers and proceedings
2oral presentations at IEA 2024 and IISE 2025
4research labs across the US and China

Publications

* equal contribution

Papers on how AI behaves in clinical and biological settings.

Journal articles

  1. JMIR

    Advanced Artificial Intelligence Systems in Assisted Reproductive Technology: Evidence of an Alignment Paradox

    Dou Liu, Ying Long, Sophia Zuoqiu, Di Liu, Kang Li, Yiting Lin, Hanyi Liu, Tian Tang, Rong Yin

    Journal of Medical Internet Research

    DOI ↗
  2. JMIR

    Reliability of Large Language Model–Generated Clinical Reasoning in Assisted Reproductive Technology: Blinded Comparative Evaluation Study

    Dou Liu, Ying Long, Sophia Zuoqiu, Di Liu, Kang Li, Yiting Lin, Hanyi Liu, Tian Tang, Rong Yin

    Journal of Medical Internet Research

    DOI ↗
  3. IISE THSE

    Automated Procedural Analysis via Video-Language Models for Scalable Training of Fundamental Nursing Skills

    S. Chang*, Dou Liu*, R. Tian, K. L. Swartzell, S. L. Klingler, A. M. Nagle, N. Kong

    IISE Transactions on Health Systems Engineering

    DOI ↗

Conference proceedings

  1. IISE 2025 · Oral

    Evaluating the Feasibility and Accuracy of Large Language Models for Medical History-Taking in Obstetrics and Gynecology

    Dou Liu, Ying Long, Sophia Zuoqiu, Tian Tang, Rong Yin

    IISE Annual Conference & Expo 2025

    DOI ↗
  2. IEA 2024 · Oral

    Evaluating the Application of ChatGPT in Outpatient Triage Guidance: A Comparative Study

    Dou Liu, Ying Han, Xiandi Wang, Xiaomei Tan, Di Liu, Guangwu Qian, Kang Li, Dan Pu, Rong Yin

    22nd Triennial Congress of the International Ergonomics Association

    DOI ↗

Experience

Research across genomics, clinical AI and health operations.

  1. 2026 – now

    Graduate Research Assistant

    Xia Lab, The University of Texas at Austin · Prof. Zheng Xia

    LLM multi-agent systems for cancer single-cell analysis, cell-type annotation and oncogenic gene discovery, with an explainable reasoning layer.

    Austin, TX
  2. 2025 – 2026

    Co-Founder & CTO

    ArteMEDs

    Led technical development of an AI-enabled triage and consultation-support platform: system architecture, prototype deployment and clinical knowledge-base integration.

    Munich, Germany
  3. 2025 – 2026

    Continuous Improvement Consultant

    Michigan Medicine, University of Michigan

    Cluster-based integer programming for consult routing under spatial, workload and capacity constraints, targeting a 20% reduction in consultation delays.

    Ann Arbor, MI
  4. 2025 – 2026

    Undergraduate Research Assistant

    Human Betterment Analytics Lab, Purdue University · Prof. Nan Kong

    Video-language models that assess fundamental nursing skills, trained with a hierarchical, coarse-to-fine framework and validated with the Nursing Center.

    West Lafayette, IN
  5. 2024 – 2026

    Undergraduate Research Assistant

    West China Biomedical Big Data Center · Prof. Kang Li

    Multi-hospital EHR pipelines for cleaning, de-identification and integration, used to model treatment pathways in reproductive medicine.

    Chengdu, China
  6. 2022 – 2026

    Undergraduate Research Assistant

    Human-Centered AI Lab, Pittsburgh Institute – Sichuan University · Prof. Rong Yin, Prof. Kang Li

    Built ARTLLM, measured the Alignment Paradox, and evaluated AI for outpatient triage and history-taking. Mentored three junior research assistants.

    Chengdu, China
  7. 2025 – 2026

    Undergraduate Teaching Assistant, FIN 300

    Ross School of Business, University of Michigan

    Office hours, review materials and grading for a 103-student undergraduate finance course.

    Ann Arbor, MI

EarlierStrategic Investment Analyst Intern, Hangzhou Jintou Asset Management (2024) · Asset Management Assistant Intern, Everbright Securities (2023)

Education

  1. 2026 – 2031

    Ph.D., Biomedical Engineering

    The University of Texas at Austin

    Computational Biomedical Engineering · Advisor Prof. Zheng Xia · Expected May 2031

    Austin, TX
  2. 2024 – 2026

    B.Eng., Industrial and Operations Engineering

    University of Michigan

    Financial Engineering · Minor in Business Administration · GPA 3.96 / 4.00

    Ann Arbor, MI
  3. 2022 – 2026

    B.Eng., Industrial Engineering

    Sichuan University

    GPA 3.94 / 4.00

    Chengdu, China

Honors

  • James B. Angell ScholarUniversity of Michigan
  • University HonorsUniversity of Michigan
  • McLane Family Endowed Fund2025Industrial and Operations Engineering, University of Michigan
  • Rasch Scholarship2024Industrial and Operations Engineering, University of Michigan
  • Dean’s List2024–2026College of Engineering, University of Michigan
  • Student of the Year, Highest Honor2024Sichuan University
  • Dean’s List2022–2024Sichuan University

President, Eta Omega Chi business fraternity (2024–2026) · Member, Alpha Pi Mu · Student member, IISE and INFORMS

Contact

Let’s talk cells, agents and evidence.

I’m open to collaborations on single-cell analysis, LLM agents and scientific reasoning. Email is the fastest way to reach me.

dnsliu@utexas.edu
Google Scholar ↗

Austin, Texas · Central Time