dida developed a custom, modular AI solution on Google Cloud that automates Enpal's manual sales quoting process. They implemented a machine learning pipeline that collected rooftop images via the Google Maps Platform API and stored them in Cloud Storage, then trained a robust ML model on GPU-enabled Compute Engine using projective geometry to accurately estimate roof angles. The solution streamlined Enpal’s sales workflow by integrating continuous development with Cloud Build and orchestrating model operations via Vertex AI and TensorBoard, automating complex measurements and quote generation tasks.