Email: umer.gupta152@gmail.com
Resume: PDF

Research interests

Graph neural networks, geometric deep learning, generative modelling, and ML for geospatial.

Current research (ETH Zürich)

Research collaborator working on graph based generative models for binary geometric trees like biological trees and dendritic arbours.

• Autoregressive leaf expansion + conformation prediction with a joint diffusion process
• SO(2)-equivariant EGNN variant using relative frames

Applied ML

• Monitoring nature restoration projects in the UK using geospatial models
• Reducing wastage in near-field mining exploration with 3D voxel models
• Supporting work in sustainable protein synthesis by optimising apparutus calibration using ML

Education

MSc Data Science (Distinction), University of Edinburgh
BSc (Hons) Mathematics, Delhi University