MLWatch

I'm Ramjee co-founder of Dblue along with @rajeshhegde At Dblue we focus on the Engineering side of Machine Learning, enabling Data Scientists and AI/ML researchers to focus on solving business problems and not worry about tech. plumbing. Unlike software, ML systems are built on data, data for most systems changes over a period of time. This means without a line of code change anywhere, your models will start performing badly because the underneath data distribution changed. From our 25+years background building engineering systems at scale we realize the importance of good data and constant monitoring of the same feeding into ML models, in ensuring good performance. With MLWatch we bring to you the transparency and visibility needed to understand your model behavior. MLWatch is monitoring platform to measure model performance in real-time. * Platform, tool, framework agnostic. * Logging, analysis of feature, prediction, feedback. * Feature Statistics * Model prediction, performance metrics * Deep dive explanation into every prediction * Data drift against baseline data * Data Violations (missing data, new data, schema violations etc.) * Individuals/Teams and Role based access. We are working hard to add automatic anomaly detection, alerts and more in the next few weeks. Now get the best in class Data, ML model monitoring solution with 5 lines of code. Please email me at rg@dblue.ai if you want to chat over the challenges you are facing.

Congratulation @gantir on the product launch

Great congrats @gantir on the product launch

This looks cool. What's the value prop for an engineering team to use your product vs. going with tools provided by AWS, Azure, or Google Cloud - assuming many teams are using those services already to run at least inference/serving?