Engineering for Science. Accelerating Machine Learning in Ring with AWS SageMaker

AI is on everyone's mind, whether ML, DL, NLP, DS, LLMs or CV. Researchers and ML Engineers, while eager to try new ideas and develop innovative solutions quickly, also need to keep in mind data privacy and security. So how can we move fast while maintaining customer trust? In Ring, we develop various ML solutions facing these challenges.

This presentation describes our ongoing journey in tackling them. We will talk about how we leverage AWS SageMaker to vend private experiment environments for researchers and speak to the challenges that come with that. We will talk about how we help researchers train models at scale and how we feature their solutions so that other Ring engineers can discover and use them. We aim to increase collaboration and democratise ML with our solutions. To move fast and support multiple research teams, we use AWS and infrastructure-as-code (IaC) with AWS CDK and CI/CD pipelines.

Meet the speaker

Daniel Vecliuc
Daniel Vecliuc Software Engineer at Amazon

Daniel is a software engineer at Ring, an Amazon company, the smart home security division aiming to make neighborhoods safer. As part of the Ring AI division, Daniel is working to improve the research process focusing on creating a privacy-oriented centralized Data Platform for all the existing data in Ring. Before joining Ring, Daniel gained experience on multiple business areas such online threat security, medical, municipal facilities, and sales & taxes. Additionally, he contributes as a laboratory assistant and pursues his Ph.D. in the field of Multi-Agent Systems at the Automatic Control and Computer Engineering faculty in Iasi.