Secure note sharing with decentralized storage and incentive design.
Prototyped a Web3 marketplace where students could publish, verify, and exchange high-quality study notes through IPFS-backed storage and Solidity-based reward mechanisms.
Links: IPFS ↗ Solidity ↗
Smart-health food identification and nutrition estimation system.
Built a lightweight object-detection pipeline that identifies foods and estimates nutritional values, demonstrating practical diet-tracking performance for real-world use.
Links: IEEE Xplore ↗
Campus-specific monitoring for dorm laundry availability.
Built a vibration-sensor device and web dashboard that tracks machine usage patterns in campus laundry rooms, making wait times visible before students walk across the dorm.
Machine learning pipeline for telescope readout noise.
Designed a DBSCAN-PCA workflow for identifying and filtering cosmic ray artifacts in telescope detector readouts, improving runtime by 10x while preserving interpretable clustering behavior.
Pipeline
jwst-noise
Real-time semantic segmentation for rover navigation.
Created a lightweight MobileNetV3 and DeepLabV3 architecture for rover terrain parsing, tuned for low-latency edge deployment on Jetson Nano hardware.
Links: Jetson Nano ↗
AI-powered tutoring platform for autistic students in grades 2-10.
Developed a personalized learning platform deployed across 16 Title I schools, pitched to 14 Fortune 500 firms, and secured $12K in funding through the Congressional App Challenge ecosystem.