Technology Spring Intern

Selected for Amazon’s Spring Insight Programme, working in a team on a predictive modelling problem and deploying it on AWS infrastructure.

Predictive modelling challenge

My team ranked 2nd of 20 in the application development challenge. We built a system to analyse film scripts and predict audience reception, aimed at informing production funding decisions before the money is committed: a decision-optimisation problem where being wrong is expensive in one direction and merely disappointing in the other.

Deployment and monitoring

I deployed a Flask application on EC2, configuring security groups, REST APIs and CloudWatch monitoring for a live service, and worked with SageMaker endpoints for model deployment. Standing up a service that is actually reachable, actually locked down, and actually observable is a different skill from making the model work locally.