I am a full-stack developer with a passion for architecting and building robust applications.
I’m also deeply engaged in machine learning, particularly in the areas of object detection, segmentation,
and computer vision.
My work and curiosity often lead me to explore the applications of machine learning in climate change,
remote sensing, and healthcare — with a focus on building end-to-end pipelines and addressing
software architecture challenges. I also keep close eye on Explainable AI, Domain Generalization and Data Privacy.
Outside of technology, I enjoy learning through reading and traveling, and I find peace and inspiration in nature.
Publications
- Co-author, "Automatic Suspicious Behavior Detection from a Small Bootstrap Set," Proceedings of the International Conference on Computer Vision Applications (VISAPP), 2012.
- Co-author, "Incremental Behavior Modeling and Suspicious Activity Detection," published in the Journal of Pattern Recognition, 2013.
Presentations
- Program/Open Data/TPG87N-Data use and reuse: Unlocking the Potential of Frictionless Data for Wikidata, 2023
Projects
In this project, we are currently working on building pipeline for disaster monitoring and real-time data analysis features for post-disaster impact mitigation.
This was a group project as part of NMA(Neuromatch Academy) course, where we replicated the work of Kay and colleagues by applying machine learning techniques to train a receptive field encoding model to predict brain activity.
We trained various object detection models including RCNN and U-NET to analyze satellite images for damage caused by army worms.
Frictionless-py is a data management framework for Python that provides functionality to describe, extract, validate, and transform tabular data.
Open Data Editor is an open source tool for non-technical data users to clean and process data before publishing.
Ckan is an open-source DMS(data management system) for powering data hubs and data portals.