About

I am a final-year Ph.D. candidate in Computer Science at the University of Illinois Urbana-Champaign, co-advised by Professors Vikram Adve and Yu-Xiong Wang. I work on visual and multimodal learning for dynamic real-world environments. My research explores how AI systems can adapt to new categories and environments, transfer what they learn across models and tasks, and ground their predictions in reliable external knowledge.

I am particularly interested in building systems that move beyond closed-world benchmarks and remain useful when deployed in the real world. My work spans multimodal foundation models, open-world detection, 3D perception, and knowledge-grounded reasoning, with applications ranging from agricultural robotics and wildlife understanding to retrieval-augmented multimodal models.

Before starting my Ph.D., I spent four years in the industry as a software engineer at Microsoft, where I worked on large-scale production systems as part of the Azure Backup team.

Experience

Visiting Scholar Bonn, Germany

University of Bonn May – Aug 2025

Advisor: Prof. Cyrill Stachniss

Developed a structure-augmented 3D perception framework for robotic pruning, along with a large-scale annotated LiDAR dataset of real orchard trees.

Research Intern, Research for Industry Redmond, Washington

Microsoft Research May – Aug 2024

Mentors: Dr. Emre Kıcıman and Dr. Ranveer Chandra

Built a learnable retrieval framework that adapts large multimodal models to new tasks without fine-tuning, transferring across models including GPT-4o and Gemini.

Software Engineering Intern, Amazon Care Seattle, WA

Amazon Jun – Aug 2021

Built a serverless publisher–subscriber system on AWS Lambda for Amazon Care’s Payments service.

Software Engineer I & II, Azure Backup Hyderabad, India

Microsoft Jul 2016 – Aug 2020

Shipped cross-region disaster recovery for Azure VM, SQL, and SAP HANA backups to general availability, and built the analytics pipeline behind Azure Backup’s monitoring dashboards.

Publications

Teaser for orchard pruning paper

Structure-Augmented Learning from 3D Point Clouds for Orchard Pruning

Under Review

Garvita Allabadi, Matteo Sodano, Elias Ariel Marks, Gianmarco Roggiolani, Jens Behley, Yu-Xiong Wang, Vikram Adve, Cyrill Stachniss

GRIP paper thumbnail

GRIP: Feedback-Guided Prompt Retrieval for Large Multimodal Models

Under Review

Garvita Allabadi, Matteo Sodano, Roberto Estevão, Yu-Xiong Wang, Vikram Adve, Emre Kıcıman, Ranveer Chandra

SAM in the wild paper thumbnail

Learning to Detect Novel Animals with SAM in the Wild

International Journal of Computer Vision (IJCV), 2024

Garvita Allabadi, Ana Lucic, Yu-Xiong Wang, Vikram Adve

Open-world semi-supervised detection thumbnail

Generalized Open-World Semi-Supervised Object Detection

NeurIPS Workshop on Open-World Agents, 2024

Garvita Allabadi, Ana Lucic, Siddarth Aananth, Tiffany Yang, Yu-Xiong Wang, Vikram Adve

CV4Animals workshop paper thumbnail

From N to N+1: Learning to Detect Novel Animals with SAM

CVPR Workshop on Computer Vision for Animal Behavior Tracking and Modeling, 2023

Garvita Allabadi, Ana Lucic, Yu-Xiong Wang, Vikram Adve

smol LoRa soil moisture thumbnail

smol: Sensing Soil Moisture using LoRa

MobiCom Workshop on No Power and Low Power Internet of Things, 2021

Garvita Allabadi*, Daniel Kiv*, Berkay Kaplan, Robin Kravets

METIS runtime verification thumbnail

METIS: Resource and Context-Aware Monitoring of Finite State Properties

International Conference on Runtime Verification (RV), 2018 · Best Paper Award

Garvita Allabadi, Aritra Dhar, Ambreen Bashir, Rahul Purandare

* indicates equal contribution.

Awards

Service

Teaching

  • Teaching Assistant

    Computer Architecture (CS 233)

    Aug – Dec 2026

  • Lead Teaching Assistant

    Data Visualization (CS 416)

    May – Aug 2026

  • Teaching Assistant

    Introduction to Computer Science (CS 125)

    Aug – Dec 2020