Monitoring AI-Powered Real-Time Image Analysis w/ Grafana’s Observability Stack

Join the DC Grafana & Friends meetup group for a presentation from Christopher R. Field of Theia Scientific, Kevin G. Field of Theia Scientific and University of Michigan - Ann Arbor, and Mikhail Volkov of Volkov Labs.

Complementary snacks and soft drinks will be on hand. We look forward to seeing you there!

Theia Scientific, LLC in partnership with Volkov Labs, LLC have developed a hardware-software platform for real-time analysis of scientific and engineering images using a combination of Grafana’s observability stack, Artificial Intelligence (AI), Internet-of-Things (IoT) hardware, and distributed computing libraries.

Here, we will present the challenges faced with AI model deployment in network and resource-constrained environments, such as those found in academic, government, and industrial Research & Development (R&D) laboratories. We used Grafana’s observability stack to spot problems, identify bottlenecks, and provide solutions to these challenges, which yielded a 80% in time and cost over manual post-acquisition image analysis and quantitation techniques for scientists and engineers.

Specifically, we will discuss leveraging custom data source and panel plugin development, integrating a variety of native and community-supported data sources, including Prometheus, HTTP APIs, WebSockets, PostgreSQL, and Environment, scaling the platform to heterogenous ad-hoc clusters for dynamic resource management, and creating beautiful dashboards for science and engineering.

A demonstration and application walkthrough will be provided as part of the discussion.