Brian Wylie

Models get created on Sandbox, then then get promoted to Prod. Outline -How to models get created on Sandbox? - How to models get moved (S3 sync.. shrug) - Once...

Let's organize/combine the current endpoint inference/prediction methods. Basically there will be one method that can take multiple args: `endpoint.inference()` - has a capture flag (defaulted to False) - can take...

Lets put together a set of Architecture Documents for SageWorks: - Top Level Page - Artifacts - DataLoader - DataSource - FeatureSet - Model - Endpoint - Transforms - DataSource...

Let's deploy our Dashboard on the SCP Sandbox account

RedisCacheMonitor Class - Spiders the Redis Cache - Identifies OLD entries - Identifies LARGE entries

SageMaker Pipeline have lots of benefits, AWS managed, scalable, AWS invokable, etc... so as part of the new API work lets have a new SageWorks Pipeline class that basically takes...

So the new config manager will store SageWorks config for like API Keys, etc.. but it will also store Feature Flags that can be turned on and off. We need...

AWS Canvas at top of pyramid, AWS services at bottom, big gap in between, which is filled by SageWorks. - Medium Granularity - Legos - Workbench

Right now in the connectors/endpoint.py class we're pulling in details at the same time as pulling in summary.. lets split those two apart so that requests are 'lighter'.

Right now using the AWSServiceBroker is WAY too confusing for most folks. So for Plugins and just general code usability lets make a MetaDataManager Class that simplifies the retrieval of...