Welcome to the Dataiku community.
Auto ML jobs are usually computationally complex. Depending on how much computational power the computer you are running dss on has, you may not want to increase the number of concurrent jobs. This may actually slow things down because you don’t have enough CPU cores or RAM memory. Note the resources we are talking about are not your local computer but the server(s) on which DSS is running.
I’d take a look at top, or htop or one of the other tasks manager on the computer running these jobs to see if you have significant unused system resources. This could be more complicated to evaluate if you are doing your compute on kubernetes.
If you do find that you have a bunch of unused resources you might consider increasing the concurrent limits. See the documentation pages.