I use loopback to submit a task to a worker queue and for the actual worker infrastructure i'm using a hosted solution known as Iron.io
Here's a brief article on how I switched from Parse background jobs to Iron.io:
1) I post a job to the IronWorker queue
2) The IronWorker uses loopback models and strongloop remoting to run code that is syntactically very similar to the code written on the loopback side ... it basically crunches data then saves the data using loopback api or makes some reads if it needs to etc
3) on my loopback UI client, folks can click a job row to submit a task update request to the
iron.io api or simply check the loopback model to see if the results have maybe been saved back by the IronWorker already
Also my worker code is open source so you can use that for reference and mimic it if you'd like: