To push a model into production, there are additional concerns which the tools in the versioning, deployment and release space aim to solve. This includes obtaining adequate infrastructure to run the model reliably and facilitating easy model release or rollback. Solutions in the MLOps space includes Kubeflow, Pachyderm and Algorithmia. - Source: dev.to / over 2 years ago

And for enterprises that want to do the same with ML you can use algorithmia.com. Source: over 2 years ago

Algorithmia advertises themselves as an MLops platform for data scientists, and they provide an easy way to host models on a scalable REST API. Source: over 2 years ago

Seems similar to https://algorithmia.com. Source: over 2 years ago

Algorithmia.com — Host algorithms for free. Includes free monthly allowance for running algorithms. Now with CLI support. - Source: dev.to / almost 3 years ago

Https://getporter.org/ https://getporter.dev/ One of you is going to have to rename yourselves... - Source: Hacker News / over 1 year ago

Porter - a fully-managed PaaS that lets teams automate DevOps. The free basic tier for porter cloud offers management of 1 cluster with up to 10 vCPU and 20 GB memory. - Source: dev.to / over 1 year ago

There are some YC startups (AtomizedHq.com and getporter.dev) that are doing really interesting things with cross-cloud K8S deployments (more like heroku). These are all different bits of the serverless microservices scaling puzzle. We are a long way off but trying to think long term, even as a 2 person alpha prototype :). Source: almost 3 years ago

But then I saw a YC startup called Porter (https://getporter.dev) that made getting the cluster set up and deploying the apps from Heroku on AWS EKS a piece of cake. It's really great. There is another YC startup called Atomized (https://atomizedhq.com) that I've been looking at that's also really great. They are both worth checking out, and the teams from both are super-responsive. Source: almost 3 years ago