3 ms·
Data Distribution Shifts and Monitoring
- mrbuzzinfrog 5y agoA very important topic. Specially if you are into MLOps! If your product is mainly powered by AI, and you found "AI-product-fit", then this topic should be something you bring up to the AI team tomorrow or to the infra team. AI models are powerful yet sensitive, and how new samples (out-of-dist) could affect them is hard to anticipate. The other day I came across https://www.raymon.ai/ https://www.raymon.ai/, "Observability for AI systems ". I'vent tried Raymon yet, but looks promising.