2023 · Machine learning
Earthquake building-damage detection on AWS SageMaker
a national government ministry · Machine learning & cloud infrastructure engineer
The problem
After an earthquake, a ministry gets aerial images within hours but has no fast way to turn thousands of image tiles into a map of damaged buildings.
What I built
I deployed a custom computer vision model as a service on AWS and built the pipeline that turns its detections into map data, all provisioned as code.
The result
Responders get a map of likely damage instead of reviewing thousands of image tiles by hand.
Main tools
- Amazon SageMaker
- AWS Lambda
- Amazon S3
- Terraform
- Detectron2
This is the short version. The full case study is available on request.