Re:Michelangelo: Uber’s Machine Learning Platform
by rasmi on Wednesday September 06, 2017 @10:32AM
(#15183976)
Disclosure: I work for Google. Google Cloud does this with Dataflow and ML Engine in a fairly simple way, and the internals are open sourced as Apache Beam and TensorFlow so you can migrate to other platforms if you want.
Re:Michelangelo: Uber’s Machine Learning Platform
by mayank on Wednesday September 06, 2017 @02:19AM
(#15181594)
Not quite there, but parts are:
* You can send your firehose to BigQuery, and write Javascript UDFs to transform the data.
* You can use a service like DataRobot to automatically train a bunch of models.
* You can deploy the trained models using a serverless platform like Lambda.
Nothing that's one-size-fits-all, but watch out for trends in that direction as more ML workflows get standardized.
Re:Michelangelo: Uber’s Machine Learning Platform
by wmeddie on Wednesday September 06, 2017 @08:14AM
(#15182985)
Skymind's SKIL does this. (I work on the team.) You don't have to send it your data though, you can just install it inside your system and have it connect to the data source or stream to train models (optionally doing hyper-parameter search), deploy them to REST servers and monitor their performance.
Re:Michelangelo: Uber’s Machine Learning Platform
by mobileexpert on Wednesday September 06, 2017 @01:49AM
(#15181499)
Azure ML
Re:Michelangelo: Uber’s Machine Learning Platform
by ANaimi on Wednesday September 06, 2017 @06:41AM
(#15182591)
Algorithmia Enterprise (algorithmia.com/enterprise). Inference- and deep learning-centric. (I'm part of the team).
Re:Michelangelo: Uber’s Machine Learning Platform
by romankolpak on Wednesday September 06, 2017 @02:32AM
(#15181637)
You should take a look at DataRobot, it does most of the things you've listed.
Re:Michelangelo: Uber’s Machine Learning Platform
by pplonski86 on Wednesday September 06, 2017 @06:53PM
(#15187794)
I wrote MLJAR (https://mljar.com) that is doing model and parameters search - and it helps me a lot with ML projects.
That is, suppose I'm willing to do the following:
* Send said service a firehose of data
* Write code for feature selection, data cleanup, transformations, etc. -- i.e. the sort of code you'd expect to find in a data scientist's Jupyter notebook.
* Specify the type of ML model and parameters I'd like to experiment with.
Is there a service that handles data warehousing, serving up predictions via a REST API, deployment of new models, reporting model performance over time, and general scaling issues? I know it's possible to set up a large portion of this using AWS services, but ideally, it'd be nice to have something with a Heroku-like ease-of-use, but for ML.