Navodita Mathur

Navodita Mathur

AI Research intern at Dugree

Navodita Mathur

AI Research Intern, Machine learning Engineer
Location: San Mateo, CA
Contact: +1(412) 954-7877
Email: navoditamathur1998@gmail.com
Github: Navoditamathur

About Me

I am an AI Research Intern at Dugree with over three years of experience in software development. My expertise lies at the intersection of Computer Vision, Satellite Imagery, and Remote Sensing, where I’ve contributed to innovative solutions across cutting-edge projects. I am driven by a deep passion for utilizing technology to better understand and safeguard biodiversity, which fuels my current research focus in Multi-Modal Machine Learning and Artificial Intelligence (AI).

I am deeply interested in the emerging field of imageomics, which applies computer vision to biological data to extract species-level insights from images. Currently, I am currently working on integrating diverse modalities—such as language, camera traps, bioacoustics, and environmental DNA (eDNA) to develop a holistic ecosystem monitoring framework.

My aim is to leverage these advanced technologies to monitor, predict, and protect ecosystems, contributing to global conservation efforts. By combining the latest in AI, machine learning, and data analysis, I aspire to create innovative tools that support biodiversity conservation and promote sustainable environmental practices.

Key Research Projects

  • Brain Tumor Image Segmentation: Developed an advanced classification model utilizing Swin transformers and attention mechanisms to accurately segment and classify brain tumors.

  • Road Crack Image Segmentation: Implemented a novel approach using discrete wavelet transforms and self-attention mechanisms to detect and segment road cracks with high precision.

Further details in portfolio

Research Experience

As a Machine Learning Engineer (Volunteer) at Omdena, I have had the opportunity to work on projects that address critical social and environmental issues through the application of advanced machine learning techniques. My role involves developing comprehensive end-to-end pipelines that are integral to the success of these projects, allowing for seamless data processing, model training, and deployment. My work at Omdena has enabled me to leverage my expertise in computer vision, remote sensing, and machine learning to tackle real-world challenges and deliver impactful solutions. Two of the key projects I have been deeply involved in are focused on Yield Prediction and Flood Forecasting. These projects exemplify how technology can be harnessed to improve agricultural productivity, enhance disaster preparedness, and ultimately contribute to global sustainability. I provide a detailed overview of each project and its significance in research tab.

Explore my GitHub repository for a comprehensive list of notebooks and code related to my ecological AI experiments

For my detailed implementations of research papers, please refer to this repo.