AI/ML Open-Source solutions are becoming more prevalent with the increased availability of data along with faster and low-cost processing compute. Neethu, alongside her team, created an AI/ML Data pipeline Processing infrastructure to enable healthcare use cases with open source EdgeX Foundries microservices that automatically detect, manage, and process images received from OEM equipment. This project enabled managing AI/ML pipelines for image processing and automating image comparisons, and developing an adaptable solution for different use cases and settings, leveraging a containerized microservices based architecture and various communication APIs & messaging bus including MQTT. This project included interesting challenges such as a distributed deployment at the edge scenario where some services ran on windows-based OEM equipment, while others ran on a Linux Edge box, timing-dependent issues leading to a lack of idempotency in integration tests, and discussions around the idea of adding wait strategy for dependent services to be up and ready before accepting TCP connections. This session will cover the architecture & design, with integration of EdgeX features to develop this solution for Edge devices along with the learnings from this open-source project. Open-Source code for setup, install, and execution of software, with complete developer documentation – https://github.com/intel/AiCSD Co-sponsored by: Neethu Elizabeth Simon Speaker(s): Neethu Elizabeth Simon Virtual: https://events.vtools.ieee.org/m/377039
This event has passed.