3 ms·
I'm working on a new Machine Learning Platform as a Service offering. We will be providing an easy to use, API driven platform for quickly and easily spinning
by mindcrime 11y ago
I'm working on a new Machine Learning Platform as a Service offering. We will be providing an easy to use, API driven platform for quickly and easily spinning up an environment for ML based on Spark and (in the first version) SystemML. Eventually we'll add support for working with TensorFlow, Warp-CTC, etc. I also envision having APIs for OCR, NLP, etc. as we develop.
One thing about this: everything will be Open Source. Obviously all of the "back end stuff" (Spark, SystemML, TensorFlow, etc.) is already open source, but even our provisioning code and API's will be OSS as well. The idea is to put the customer in control. So if you don't like our service, or decide you can do it cheaper running your own hardware on premises, whatever, you can deploy the entire stack yourself.
There's nothing at the URL right now but a (very) primitive landing page, but if any of you guys want to take a look, here ya go:
http://neuralobjects.com http://neuralobjects.com
Everything is hidden behind a (temporary) login page right now, since this isn't really public yet. The login is 'user123' with password 'realitybomb40'
Also, ignore the video that's there, it's a placeholder to mark out where our explainer video will go when it's recorded.
We plan to eventually offer both "pre canned" routines that you can use out of the box as well as the ability to deploy your own models. There a lot of other stuff on the roadmap as well, but it's probably too early to talk about any of it just yet.