Deep Learning for Satellite Contrail Detection

Job ID: 38912050

Budget: $30 – $250 USD

I need a Python expert to set up and run a deep learning project on my computer for detecting satellite contrails using the GOES-16 satellite data. The project will use the UNET model and involve multiple steps from dataset preparation to model evaluation and inference.

Key Requirements:
- Write and execute code for three different deep learning models to compare their performance.
- Setup the necessary environment on my computer to run the code on my GPU.
- Create a detection code for uploading and testing images.
- Provide a comprehensive explanation of the code and the workflow.

The project is based on the Google Research competition: https://www.kaggle.com/competitions/google-research-identify-contrails-reduce-global-warming. It involves:
- Generating pipeline code from the dataset.
- Sequentially performing setup, test training, evaluation, and inference coding for each model.
- Using Python for the entire process including the environment setup script.

Ideal skills and experience:
- Proficiency in Python and deep learning frameworks, particularly TensorFlow.
- Experience with satellite data analysis and contrail detection.
- Familiarity with the UNET model.
- Ability to explain complex code in a simple manner.