
CrisisLens
1st Place, Hacklytics Databricks × UN Challenge. AI-powered crisis forecasting to help prioritize humanitarian aid.
Here are some representative projects I have worked on, across machine learning, robotics, computer vision, and the web. Each project includes the problem, my approach, and what I learned.
You can find more of my work on GitHub and my technical writing on the blog.
14 projects

1st Place, Hacklytics Databricks × UN Challenge. AI-powered crisis forecasting to help prioritize humanitarian aid.

Most Innovative Hack among 280+ projects. Voice-controlled navigation with LiDAR mapping and ROS 2.

1st Place Hackathon Winner. Four AI agents streamline personal injury cases through document processing, case analysis, and client communications.

Short-form financial education with AI content screening and Solana token rewards. Built at SwampHacks XI.

Predicting pasture biomass from images with a four-model vision ensemble and stacked regression.

AI-assisted mutual aid that connects local offers, requests, and volunteers around urgent community needs.

An EEG-based brain–computer interface connecting Muse headband signals to real-time drone commands.

Exploring RNA structure prediction with sequence features, evolutionary context, and a dual-tower Transformer.

Gesture recognition from motion and time-of-flight sensors using dedicated encoders and temporal fusion.

Voice-assisted ordering and vision-guided burger retrieval with a six-axis robotic arm.

Artist discovery through audio-feature clustering, similarity ranking, and Spotify track previews.

Exploring historical Chicago crime records with temporal features, classification, coordinate regression, and map visualizations.

Comparing Random Forest, support vector regression, and linear regression on market data with streamed model results.

Exploring monocular camera tracking, feature matching, and relative-depth reconstruction for drone navigation.