Introduction & Architecture Question-Mithul Krishna

[MithulKRS]

Hi, I am Mithul Krishna Suresh, a 2nd-year CS student from Maulana Azad National Institute of Technology (MANIT) in Bhopal. I am experienced in Python, deep learning (CNNs), time-series forecasting, and external API integration.

This summer, I am looking forward to contributing to Project Idea [3] Alaska Wildfire Prediction Using Satellite Imagery, because it aligns perfectly with my previous experience in building multimodal systems and handling temporal weather data.

My sample contributions relevant to this project can be found here:

https://github.com/MithulKRS/Hyper-local-Demand-Forecasting-Inventory-Optimization-Engine : Integrated historical data with real-time weather APIs to predict demand spikes using TimeSeriesSplit validation.

https://github.com/MithulKRS/AI-Plant-Disease-Detective-Pathologist : Engineered a multimodal diagnostic system using custom CNNs and LLMs, containerized via Docker.

Question regarding the dataset preprocessing pipeline:
For the initial satellite data preprocessing (especially with Sentinel-2 and Sentinel-1 SAR), do you recommend focusing on local tools like rasterio or xarray? Or is it better to use cloud-based platforms like Google Earth Engine for things like cloud removal and cropping?

Best Regards,
Mithul Krishna Suresh

[pradeeban]

@MithulKRS Pls ask this question in the relevant GitHub discussion channel of idea 3 as listed in our ideas page: https://github.com/YaliWang2019/AK-Satellite-Imagery-Wildfire-Prediction/discussions