ClimateHack.ai 2022

• Ranked 6th out of the 25 top universities from the UK, US and Canada.
• Developed a model in PyTorch to predict solar photovoltaic power production using satellite imagery.
• Increased the receptive field of the sequence-to-sequence model using UNet-inspired components and improved the gradient flow of the network by making use of residual connections, which led to a 10% increase in the validation metric.