I turn complex problems
into intuitive products.

Carrying heavy gear across campsites, warehouses, or shopping areas is physically demanding — and existing electric carts still require manual steering with no obstacle avoidance or person-following capability.

Designed and built a vision + UWB sensor fusion mobile robot that autonomously follows a user while avoiding obstacles in real-time, integrating OAK-D perception, ROS2-based control, and differential drive motor control into a complete working prototype.

Working prototype achieved real-time person tracking, autonomous following at walking speed, and depth-based obstacle avoidance — full hardware + software integration from perception through planning to motor control.

Creating XR content requires manual 3D modeling, importing, and scene placement — a slow pipeline disconnected from the physical objects designers actually want to reference.

Built a physical copy-paste interface: point a camera at a real object, press Copy, and AI recognition instantly spawns a digital 3D representation in your XR environment.

Working research prototype demonstrating real-time physical-to-digital object transfer — combining edge AI vision, custom hardware controller, and Unity XR environment into a seamless interaction.

Snack consumption is rising while physical activity declines — but health tracking systems treat diet and activity as separate data silos, missing the relationship between the two.

Built an IoT snack box prototype with embedded sensors that logs snack events, integrates Apple Health activity data, and uses AI to generate personalized lifestyle reports connecting diet habits with physical activity.

Working proof-of-concept demonstrating end-to-end data flow from physical snack event capture through health data integration to AI-generated daily and weekly lifestyle reports.

Clinical-grade PPG monitors were bulky and uncomfortable for 24/7 wear during sleep studies.

Flexible PCB wearable with optimized sensor placement. BLE streaming to companion app with real-time visualization.

93% participant compliance (up from 61%). Data quality matched clinical benchmarks.

Operators needed weeks of training to control a 7-DOF robotic arm through CLI-based tooling.

Ergonomic physical controller with force feedback + React telemetry dashboard showing arm state in real-time.

Training time reduced from 3 weeks to 2 days. Operator error rate dropped 74%.

Agricultural monitoring required manual data collection across 500+ acres — impractical and unreliable.

LoRa mesh network with solar-powered sensor nodes. Cloud platform auto-generates actionable alerts.

100% coverage with 2-year battery life. Water usage optimized by 30%.