CV

Research CV focused on reliable machine learning, selective prediction, uncertainty quantification, distribution shift, multimodal health sensing, and medical AI.

Contact Information

Name Jaden (Minjae) Moon
Professional Title Computer Science Researcher
Email jaden.moon.27@dartmouth.edu
Location Hanover, New Hampshire

Professional Summary

Dartmouth computer science student studying reliable machine learning for health, with research in selective prediction, uncertainty quantification, distribution shift, multimodal health sensing, and medical AI.

Education

  • Dartmouth College
    B.A. in Computer Science
    Hanover, NH
    2023 –2027
    GPA: 3.92 / 4.00 Expected graduation: June 2027.
    Relevant coursework: Deep Learning, Deep Learning Robustness, Machine Learning, AI Agents, Software Design & Implementation, Linear Algebra, Differential Equations

Publications

Research Experience

  • James O. Freedman Presidential Scholar
    HealthX Lab, Dartmouth College
    Hanover, NH Advisor: Prof. Andrew T. Campbell
    Sep 2025 –Present
    • ReliaGate (accepted at WellComp 2026): Developed a post-hoc reliability router combining confidence, signal quality and modality trust, cross-modal agreement, train-standardized atypicality, and train-fitted embedding geometry to surface or withhold unchanged wearable-stress predictions.
    • Evaluated ReliaGate across WESAD, UBFC-Phys, EmpaticaE4Stress, and PhysioNetE4 using subject-disjoint folds, fold-local selection, and paired held-out-subject intervals; on UBFC-Phys, relative to the learned no-geometry routing system, increased coverage by 10.1 percentage points and reduced accepted risk by 5.6 percentage points. WESAD analyses showed favorable pooled outcomes could coexist with higher median per-subject accepted risk.
    • ReliaGate++ (ongoing follow-on work): Extending static routing into validation-locked temporal emission control and auditing cue precision, false stress and reassurance, contradiction and flip rates, subject reach, and episode coverage.
    • INTERSPEECH 2026: Developed a leakage-safe intervention diagnostic that freezes trained experts and fusion parameters while shuffling quality scores across held-out examples to test decision-level reliance. Across StressID and CMU-MOSEI, 25 runs per dataset showed near-zero native performance changes, while positive controls confirmed sensitivity when reliability tracked correctness.
  • Summer Undergraduate Research Fellow
    Mayo Clinic Comprehensive Cancer Center
    Phoenix, AZ Advisor: Wei Liu, Ph.D.
    Jun 2026 –Present
    • Developed a validation-locked selective-review model for quality assurance of AI-generated head-and-neck organ contours using CT, geometric, segmentation uncertainty, and native-plan dose-context evidence; classified 59.4% of 3,812 held-out contour instances as low risk at 99.91% NPV, 97.47% sensitivity, and 0.936 AUROC.
    • Built a patient-disjoint training and evaluation pipeline spanning 752 cases, 37,228 contour instances, and 5 folds with nnU-Net and SLURM; added automated checks for split leakage, provenance, and reproducibility.
  • Waterhouse Research Award Scholar
    AIM HIGH Lab, Dartmouth College
    Hanover, NH Advisor: Prof. Nicholas C. Jacobson
    Feb 2025 –Aug 2025
    • Fine-tuned LLaMA-3.1-8B-Instruct with LoRA/QLoRA for therapeutic-dialogue generation and implemented reproducible training and inference workflows under constrained compute.
    • Evaluated 1,753 generated responses using semantic metrics and structured failure analysis to identify contextual grounding and off-target generation errors.

Honors, Awards & Research Funding

  • 2026
    Mayo Clinic Summer Undergraduate Research Fellowship (SURF)
    Mayo Clinic

    $6,000 competitive summer research fellowship.

  • 2026
    Passport AI Grant
    GoPassport Education

    $2,000 external support for Mayo Clinic SURF.

  • 2026
    Best Class Project Citation
    Deep Learning Robustness (COSC 89.31), Dartmouth College

    Winter 2026.

  • 2025
    James O. Freedman Presidential Scholar
    Dartmouth College

    $3,400 research scholarship.

  • 2025
    Waterhouse Research Award
    Dartmouth College

    $6,700 summer research grant.

  • 2023–2026
    Dean's Honor List
    Dartmouth College

Skills

Methods: Selective prediction, Uncertainty estimation and calibration, Distribution-shift evaluation, Multimodal fusion, Time-series modeling, Representation analysis, Subject- and patient-disjoint evaluation
ML and Scientific Computing: Python, PyTorch, scikit-learn, Hugging Face Transformers/PEFT, NumPy, pandas, SciPy, nnU-Net
Systems and Reproducibility: Linux, SLURM, Git, AWS, Flask/FastAPI, Experiment automation, Provenance tracking, Reproducible pipelines
Signals and Data: PPG/BVP, EDA, ECG, Temperature, OpenSMILE, NeuroKit2, BioSPPy, DICOM, NIfTI

Projects

  • Local Mental Health Finder

    Built and evaluated a full-stack therapist-discovery system using symptom-based search and learning-based provider ranking under sparse, noisy inputs. Code

Leadership and Service

Republic of Korea Marine Corps (ROKMC) — Sep 2021–Mar 2023: Communications Specialist & Interpreter (Squad Leader), Hwaseong, South Korea. Led a 20-person unit and initiated “Weekly Hour,” a peer-support program designed to identify unspoken distress in a high-stigma environment. Managed headquarters networking infrastructure, resolved 150+ IT support cases monthly, and provided technical interpretation during joint U.S.–ROK exercises.
Dartmouth Tech Consulting Group — Sep 2024–Present: Led web-development and digital-strategy projects for campus organizations, delivering production systems under client and timeline constraints.
Korean Student Association — Sep 2023–Present: Organized cultural-exchange events and mentorship programs supporting international students.

Languages

Korean : Native
English : Full professional proficiency