Dylan Sherry

This is fine with me. I think its preferable to use the python built-in pickle as our default serialization strategy. I'd like this issue to track updating our pipeline and...

Pros of rolling our own: * #1936 : rather than dropping features _after_ they're generated by sklearn, if we avoid computing them in the first place, we'll save time and...

Agreed! Thanks for filing. I have been silently working on this recently :) we'll need evalml 0.24.0 to go out first. Then I'm gonna try to get [this PR](https://github.com/bentoml/BentoML/pull/1603) merged...

Blocked on bentoml doing a release. I have a notebook which is ready to go though!

Notes from discussion today: * Biggest value of having this suite on top of LG suite: immediate feedback to devs * Let's do it

For whoever picks this up: we need to confirm why we can't use woodwork with dask right now, and write up an explanation of this. Then if we need to...

@freddyaboulton great point. I agree that a) we should support setting `top_n` and other OHE params per-feature, and that b) having a "smart" value for `top_n` could be beneficial. I...

Great. I think [LightGBM](https://lightgbm.readthedocs.io/en/latest/Parameters.html#monotone_constraints) [supports this too](https://blog.datadive.net/monotonicity-constraints-in-machine-learning/). This could be a boolean option called `monotonic` specified as an automl search option. We could add a `supports_monotonic` boolean tag to pipelines...

This would be great to circle back to once the automl strategy project #272 is complete. In particular, we could define some pipelines as supporting monotonicity constraints, and if an...

Status: we're still waiting on conversation with @gsheni on this :)