cv

Professional Summary

M.S. student at Ajou University working on trustworthy and interpretable machine learning for medical imaging, with a focus on clinical faithfulness, shortcut learning, and ultrasound AI. My research examines whether high-performing medical image classifiers rely on clinically meaningful evidence, and how their reasoning can be audited without retraining.

Alongside my research, I am preparing for Mission2035 — a long-term calling to serve students in higher education as an educator and self-supporting campus missionary.

Research Interests

• Trustworthy and interpretable medical AI
• Clinical faithfulness and model auditing
• Shortcut learning and spurious correlations
• Ultrasound image analysis
• Frequency-based and graph-based representation learning

Education

M.S., AI Convergence Network, Ajou University  ·  Feb 2026 – Feb 2028 (expected) Embedded & Software Lab / MIIDS Research Center. Thesis research on clinical faithfulness auditing and shortcut analysis of ultrasound diagnostic models.

B.S., Electrical and Computer Engineering (Microdegree in Data Science & AI), Ajou University  ·  Mar 2020 – Feb 2026

Research Experience

Graduate Researcher, Embedded & Software Lab / MIIDS Research Center, Ajou University  ·  Feb 2026 – Present

• Auditing whether model explanations of ultrasound diagnostic models remain faithful to independent clinical factors rather than shortcut cues — core M.S. thesis research on trustworthy medical AI.
• Studying frequency-band and attention-based representations to analyze model reliance and explanation faithfulness across CNN-, transformer-, graph-, and wavelet-based classifiers.

Undergraduate Researcher, ITRC Project on Gallbladder Ultrasound AI (MIIDS Research Center / Ajou University Hospital)  ·  Jan 2025 – Feb 2026

• Contributed to a 3-year ITRC (IITP-funded) project on AI-based diagnosis of gallbladder disease from ultrasound, spanning clinical problem formulation, modeling, evaluation, and physician collaboration.
• First-authored a reliability requirements analysis with improved malignant-class detection sensitivity, presented at ACK 2025; continued as M.S. thesis research.
• Designed clinical-knowledge-guided modeling strategies using structured ultrasound findings (echogenicity, texture, lesion margin, wall-related features, anatomical context) as domain priors.

Research Intern, Samsung Heavy Industries Project (Embedded & Software Lab, Ajou University)  ·  Jan 2025 – Jan 2026

• Developed machine learning models for equipment fault prediction and condition monitoring in smart-factory systems.
• First-authored a monitoring-software design paper grounded in international reliability standards (KCSE 2026).

Research Intern, Department of Psychiatry, Ajou University Hospital  ·  May 2024 – Dec 2024

• Worked on a government-funded computational psychiatry project, gaining hands-on experience with IRB-approved clinical research workflows, clinical data handling, and clinician–AI collaboration.

Teaching

Teaching Assistant, Logic Circuit Experiment (논리회로실험), Dept. of Electrical and Computer Engineering, Ajou University  ·  Spring 2026

• Led weekly laboratory sessions on digital logic design — guiding students through circuit implementation and experiments, supervising lab work, and supporting evaluation.
• Student course evaluation: 4.77 / 5.0 (95.4%).

Publications

Peer-reviewed (first-authored)

• S.-H. Hwang, J.-S. Kim, M. Choi, and J.-W. Lee, “Design of Monitoring Software for Predictive Maintenance in Smart Factory,” Proc. 28th Korea Conference on Software Engineering (KCSE 2026), vol. 28, no. 1, pp. 211–218, Feb 2026. [in Korean]
• S.-H. Hwang, H. Choi, J. Kim, and J.-W. Lee, “Requirements Analysis of Data and Model for Reliability of Ultrasound-based Diagnostic Model,” Proc. Annual Conference of KIPS (ACK 2025), vol. 32, no. 2, pp. 489–490, Nov 2025. [in Korean]

Manuscripts in preparation

• Trustworthiness and clinical faithfulness of interpretable medical imaging models.
• Clinically grounded representation learning for medical ultrasound.

See the full list on the publications page.

Software Registration

A Tool for Removing Diagnostic Markers and Restoring Medical Images — Software Copyright Registration No. ASSET_0013931. Removes on-image diagnostic markers, a known source of shortcut cues in medical imaging, while preserving downstream diagnostic usability.

Ministry & Service

Leadership Team · University Group Cell Leader (목장장), Gyeongwon Church — Young Adults  ·  2025 – Present Serving and discipling university students entrusted to me — a training ground for future missions.

Campus Mission & Discipleship, JDM (Jesus Disciple Movement), Suwon  ·  2020 – 2026 · Leader 2021–2026 · Student Representative 2025 Learned that the gospel is not a program but walking long with one person — raising disciples who make disciples.

Skills

• Programming — Python, PyTorch, PyTorch Geometric, Linux, Git
• Machine Learning — CNNs, vision transformers, graph neural networks, attention mechanisms, representation analysis (incl. Riemannian geometry, hyperbolic embeddings)
• Reliability & Interpretability — shortcut analysis, frequency-band (DWT) analysis, attribution analysis, model monitoring
• Medical AI — ultrasound imaging, clinical factor analysis, image preprocessing (incl. superpixel segmentation), physician-collaborative research

Military Service

Sergeant, Republic of Korea Air Force — 15th Special Mission Wing  ·  Mar 2022 – Dec 2023 Honorably discharged.