MS Robotics student at Carnegie Mellon's Robotics Institute, working on post-training flow-matching foundation models with Reinforcement Learning in robotic manipulation tasks.
I care about two things: making robots understand the 3D world, and making them act intelligently within it. That usually means working somewhere in the intersection of:
- Robots observing the world with cameras and LiDAR
- Building structured 3D representations of scenes
- Letting vision-language or generative models reason over them
- Improving robot behavior through reinforcement learning
Robot sees world in 3D
↓
Builds rich scene representation (3DGS / NeRF / DUST3R)
↓
VLA reasons over scene + language instruction
↓
RL policy executes & improves from interaction
---
<div align="center">
<sub>Always happy to talk Machine Learning, 3D vision, or RL. Reach out anytime.</sub>
</div>