Projects

Developing a vLLM KV-cache eviction policy to reduce memory pressure during large language model inference.

Learning robotic control policies from human demonstrations using imitation learning and vision-language models.

A U-Net that removes megavoltage scatter from treatment images to support image-guided radiation therapy.

Full-stack mental wellness app offering AI-driven insights based on journal entries.

AI chatbot that teaches children about plant care as a Tamagotchi-like plant companion. Built at WildHacks.

An analysis of consent, transparency, ownership, and fairness in large-language-model training data.

A hardware-software system that detects sidewalk cracks and estimates their severity with computer vision.

Mathematical modeling of the launch conditions needed to place small-scale satellites into orbit.