Justin Minseob Seo
First-Year Ph.D. Candidate in Electrical and Computer Engineering at UC San Diego. Advancing research in multi-agentic AI efficiency, machine unlearning, and cyber-physical sensor systems.
Contact Me
AutoRPC: Rapidly Prototyping Circuit with Agentic 3D Printing
In Preparation • Targeting CHI 2027
Developing an Agentic Breadboard Circuit Digital Twin designed for cyber-physical sensor system automation. The system coordinates LLM agents to interpret natural-language design goals, generate wiring diagrams, and maintain a synchronized digital twin of the breadboard layout and 3D geometry. This links schematic-level reasoning directly with placement, routing, and multi-material geometry export.
MetaClaw Architecture & RF-Fab-DT
In Preparation
Advancing the MetaClaw system architecture to autonomously synthesize passive multi-layered metasurfaces. Engineered a deterministic JSON-based fixed lexicon to resolve inter-agent communication bottlenecks, and developed an abstract-and-reconstruct pipeline using fine-tuned small language models to translate simulation code from Ansys HFSS to NVIDIA Sionna.
AI Researcher - Cellular Layer-1 Team
Apple • Mar 2026 - Present
Specific details regarding the scope and technical contributions of this position are currently under non-disclosure and maintained in strict confidentiality.
Graduate Researcher
Prof. Xinyu Zhang's Group, UC San Diego • Mar 2026 - Present
Automating additive manufacturing workflows by integrating bambu-cli for headless slicing and direct-to-fabrication execution. Designing an in-situ mmWave radar monitoring system utilizing SAR imaging to detect internal fabrication defects, enabling closed-loop sim-to-real optimization.
Independent Computer Vision Consultant
Confidential Client, South Korea • Dec 2025 - Jan 2026
Engineered a real-time computer vision pipeline utilizing pose estimation and action recognition to track high-speed human kinematics. Developed an automated event-detection engine that translates complex physical interactions into actionable, real-time scoring data. Designed an interactive live dashboard to process video feeds and provide objective performance metrics to assist in human adjudication.
AI Researcher
Samsung Research America (SRA) • May 2025 - Present
Led development for a 6G Integrated Sensing and Communication (ISAC) project. Designed a real-time channel state information pipeline and built an interactive visualization interface for live demonstrations. Developed a multi-camera ground truth system to synchronize visual evidence with wireless sensing outputs.
Undergraduate Researcher
UC San Diego • Sep 2025 - Mar 2026
Spearheaded research on verifiable data deletion utilizing Projected Gradient Ascent. Conducted a formal geometric audit on Logistic Regression models, uncovering the "Batch Interference Effect" and establishing a 34-step Safety Window to halt unlearning before catastrophic forgetting occurred.
Sleep Apnea Detection with mmWave Radar
Samsung DA Team Collaboration
Addressed severe dataset imbalance by implementing an oversampling strategy for rare apnea events. Developed preprocessing pipelines for radar time series and optimized classification thresholds, achieving a Best Validation Accuracy of 97.33% and an F1-Score of 0.9612.
JetAuto Robotics & CSI Heatmap Interface
Samsung Research America
Integrated sensing and decision logic into a JetAuto robot platform utilizing ROS 2 and an embedded AI board for indoor perception. Developed a graphical interface to project channel state information as a dynamic heatmap of population density across office zones.
Wildfire Risk & Nutrition Dashboards
UC San Diego
Built interactive, public-facing analytical dashboards using Plotly Dash, HTML, and CSS. Focused on translating complex health and environmental data into highly readable, accessible visual layouts for non-expert audiences.
Ph.D. in Electrical and Computer Engineering
UC San Diego • Expected 2031
Research Focus: Machine Learning, Privacy, and Systems.
B.S. in Data Science
UC San Diego • March 2026
Specialization in Machine Learning and Artificial Intelligence. Coursework included Deep Learning, Machine Learning, Data Science Capstone on Machine Unlearning, Statistical Methods for Data Science, and Predictive Analytics.
- Machine Learning: Deep learning, foundation models, machine unlearning, algorithm design, PyTorch, NumPy, pandas.
- Languages & Systems: Python, C++, Java, SQL, real-time data pipelines, basic ROS 2, Linux.
- Visualization: Plotly, Dash, HTML, CSS, interactive user interface prototyping.