PRIOR (computer vision research team): Research on computer vision deep learning real-world generalization using virtual data. Integrated ShapeNet into AI2-THOR, created automated image data generation tool, and ran transfer learning using a PyTorch implementation of ResNet-50 on virtually generated data.

Incubator: Worked with company Agromo, created first product, data collection and processing web platform and app, both live.

MugScreencap