We've recently launched Drifto.

Drifto automates feature engineering over user events and transaction tables. Drifto aims to provide an automatic point-and-shoot experience: just point Drifto at any number of raw tables, and Drifto automatically combines the raw data into a large number of ML-ready features, aggregated for each ID (e.g. user) over a specified time window (e.g. weekly). Drifto supports workflows such as churn prediction, customer value estimation, anomaly detection, personalization, and demand sensing.

GitHub: https://github.com/drifto-ml/drifto Website: https://www.driftoml.com

We’d love to hear any feedback around how we can better integrate with user data pipelines or thoughts in general.

This guy gets it.

Also, recent ThinkPads do not come with good Linux support unless they are explicitly sold with Linux OS. Lenovo has different firmware for Windows vs Linux now. I recently got a Lenovo P15v Gen 2i that does not work with Linux (I ended up returning after a failed installation).

Lenovo has a >4 month backlog on Linux laptops at the moment. So we will have to make do with something else.

It is trivial that you can flip a sign on any network optimizing for something good and now it’s optimizing for something bad…

But it is still interesting that the network generalized to these toxic nerve agents without having them in the training set