I am Amandeep Kaur (Reehal), a PhD student at Arizona State University, advised by Dr. Hannah Kerner. My research focuses on understanding pretraining of large-scale Geospatial Foundational Models (GeoFMs), including the impact of pretraining dataset and the quality (or task-agnostic downstream readiness) of a pretrained model.

Previously, as part of Google Deepmind, I was working on AMED(Agricultural Monitoring and Event Detection), creating ML/DL models to generate pan-India crop covers for top-10 Indian crops using Sentinel1-2, under the guidance of Alok Talekar and Dr. Gaurav Aggarwal.

Before Google, I was at IIIT-D, pursuing my Bachelor’s in CSE where I first got into research. I had the pleasure of working with Dr Rahul Purandare in the summer of my 2nd year. Then under the guidance of Dr. Jainendra Shukla, I did an independent project where I created a system to collect audio speech samples from Indian students for our project on improving English pronunciation of Indian students. As part of my Bachelor’s Thesis, I worked on improving a head gesture guided typing system for people with limited body motion, under the supervision of Dr. Rajiv Ratn Shah.

News

2026

• First: [Internship] I’ve joined Vector Institute of AI as a Guest Researcher for the summer. I’ll be working on measuring task-agnostic downstream readiness of a pretrained model under the supervision of Dr. Evan Shelhamer.
• Update: Paper accepted to EarthVision @ CVPR. We changed no architecture. No loss function. No hyperparameters. Still performance changed dramatically. What did we change? Find out at EarthVision!

2025

• Update: [Oral] at AGU’25. Gave a talk about impact of pretraining distribution on the GeoFoundational model, SatMAE.

2024

• First: Started my PhD at Arizona State University, under the supervision of Dr. Hannah Kerner.
• First: [Oral] at EGU’24. Spoke about detecting field level crop seasons.

2023

• First: Paper accepted to AGU’23.
• First: [Oral] at AGU’23. Spoke about detecting field level crop seasons.

2022

• First: Job at Google Deepmind(previously, Google Research India). Worked with Anthrokrishi to study Indian farm fields using satellites. My google maps is always set to satellite view XD
• First: Gave a talk presenting Anthrokrishi’s work at Google for India(G4I’22).

2021

• First: Interned at Royal Bank of Scotland(Natwest Group) as a software development intern. Did a lot of Java.
• AI for Social Good: Began my two-semester Bachelor’s thesis, worked on assistive typing technology which used head gestures. Met transformers for the first time.

2020

• AI for Social Good: Woked with the Human Machine Interaction Lab, IIITD to build a system to help improve Indian students english pronunciations. Ask me about phonemes, utterances and speech in kids.

2019

• First: Research experience as part of a software development intern in the Program analysis Group, IIITD. :D