Freelance | Domain: Insurance | Zurich, Switzerland
Assessed efficient end-to-end machine learning pipelines and deployment options for Swiss digital platform for Financial Lines.
>>> AWS SageMaker, CloudFormation, Auto Scaling, Machine Learning, NLP, Computer Vision, Infrastructure as Code, Automation, Python, MLOps
• Designed approach for team that has not used Machine Learning before to take advantage of MLOps pipelines.
• Transformed ideas into concrete implementation using NLP and Computer Vision to enhance response to clients' requests.
• Improved performance and scalability, using AWS SageMaker Experiments, Debugger Insights, SageMaker Model Monitor & CloudWatch, Auto Scaling of Inference Endpoints, CI/CD.
• Automated shutdown and re-deployment of Inference Endpoints as needed using AWS CloudFormation
• Performed A/B testing using AWS SageMaker Production Variants and CodeDeploy to automate traffic shift.