projects / Research index

Selected systems

From signals
to systems.

I build across the full stack—from sensing hardware and data tools to learning systems and the experiences they make possible.

How the work connects

  1. 01
    SenseHardware & multimodal data
  2. 02
    UnderstandModels & robust inference
  3. 03
    TranslateUseful human experiences
FuSenseRing prototype worn on a finger Featured research platform
Open-source Hardware × AI

Penn State · 2025

A research-grade smart ring that fuses PPG, ECG, temperature, and contact force with temperature-adaptive learning for robust cuffless blood-pressure estimation.

4
Sensing modalities
60
Study participants
4.6 g
Complete platform
Wearable sensingSensor fusionHealth AIOpen-source hardware
Project site Read the paper

Current portfolio

Selected work.

5 systems spanning health, accessibility, robotics, and infrastructure.

AquaTag sensing devices shown at the head, wrist, and lower back 2026

Research system Penn State

AquaTag

An open-source aquatic wearable that moves across the head, wrist, and lower back while adapting its swimming analysis to each placement.

91.7% macro-F1 across placements
Aquatic wearablesEmbedded sensingWaterproof fabricationDeep learning
Explore the research
SignGlass wearable cameras used for first-person American Sign Language understanding 2026

Research system Penn State

Egocentric ASL Sensing & Learning

A connected body of wearable-vision systems for translating sign, reconstructing fast hand motion, and giving learners precise, actionable feedback.

2M+ annotated stereo frames
Egocentric visionASL3D hand reconstructionHuman-centered AI
SignGlass 3D reconstruction ASL learning
2025

Industry prototype VivoMetria

Non-invasive CGM Prototype via Near-Infrared Spectroscopy

An end-to-end feasibility prototype for continuous glucose monitoring using NIR spectroscopy, wireless acquisition, and a prototype prediction pipeline.

Live BLE-to-Android signal pipeline
Near-infrared sensingTime seriesAndroidHealthcare ML
A three-dimensional point-cloud scan of cracked pavement 2023

Research system USC Structural Health Monitoring Lab

3D Infrastructure Crack Detection

A large-scale point-cloud collection and labeling pipeline for exploring deep-learning-based crack detection across varied pavement and concrete conditions.

2,000 scans + surface images collected
Computer visionPoint cloudsDeep learning
Research lab
2023

Research system USC iLab

Interactive Mapping for Vision-based Navigation

A visual mapping workspace that made route capture, annotation, and dataset iteration faster for an autonomous navigation stack.

~20% faster data capture workflow
RoboticsComputer visionResearch tooling

Earlier explorations

Where it started.

Early work in generative tools, spatial interfaces, and team prototyping.

Team challenge · January 2022

NASA SUITS Challenge · Team Aegis

Participated with USC/UA Team Aegis in NASA's SUITS Challenge, exploring augmented-reality interfaces for future space missions.

Open source · May 2022

Synthetic Crack Generation

A MATLAB toolkit for procedurally generating realistic, varied crack structures with tree-search and computational-geometry methods.