4352 The Role: Moody's RMS Geospatial team is working on Vision Information Transformer and Large Scale Generative AI models to detect the features from remotely sensed datasets. The aim of this state-of-art AI/ML technology is to process gigabytes of satellite/aerial data to identify defined objects and patterns in a short span of time. Geospatial team is currently seeking a skilled intern who is passionate about AI/ML and willing to contribute as well as learn from AI projects. The research prospects include but are not limited to damage severity detection from wind-induced hazards' post-event images, denoising and image enhancement for better object detection, high-rise cluster detection, and change detection. Candidate during the internship will work with experienced professionals and explore cutting-edge AI technologies and algorithms in the remote sensing domain. Required Technical Skills And Educational Background Pursuing post-graduation (Masters/PhD) in Geoinformatics, Remote Sensing, Earth Sciences or Graduate in Civil Engineering, Computer Science, Environmental Science from an institute of good repute Strong mathematical concepts in machine learning, computer vision, and numerical techniques Knowledge of CNN based AI-architectures Good understanding in remote sensing technology including classification, segmentation, change detection, decorrelation algorithms for identifying the classes/objects in hyper/multi - spectral satellite/ aerial images. Strong Python programming skills and hands on exposure to Machine Learning Libraries- Scikit-Learn, Scikit-Image, Tensorflow, Keras, GluonCV and PyTorch Image processing and Geospatial Libraries - OpenCV, GDAL, Geopandas and RasterIO Knowledge of ESRI GIS software products and hands on exposure to open- source platform such as QGIS, GRASS etc. Familiarity with Google Earth Engine would be a plus. The Department / Team: The Model Development group focuses on developing industry leading stochastic catastrophe models for the quantification and management of catastrophe risk. From regional based earthquake and hurricane models to terrorism risk models, the industry, and our 400+ clients recognize RMS as the largest developer of catastrophe models in the risk market. RMS continues to drive industry innovation with the enhancement and expansion of our leading models. The Geospatial team of Model Development group is responsible for developing various geospatial layers/ products/ models required for catastrophe models.
