If I hope to enroll in a Computer Science PhD program in a top tier university (Stanford/CMU) and I currently hold a master's degree (without publication). Is a year-long Research Software development job at a reputable R&D Lab (Microsoft research lab at my country for example) or a short-term research internship at a good lab that may get my name on a paper in a good journal? Of course you may assume I cannot do both.
To expand a bit on my comments:
In PhD admission is research internship or a job [at a research lab] better?
Whatever you think gives you a better research profile at the end.
If you get a job at a strong research lab like Microsoft Research (MSR), you will typically also write strong papers, collaborate tightly with academics, and go to conferences. Hence, many of the advantages you would hope for from a research internship you would also get with a job at a good research lab. That being said, you may also get swamped with tasks more focused on "transfer" or productisation tasks, which are likely still valuable for PhD admissions, but arguably less so than actual research papers.
That being said, I am wondering whether you are (without a PhD) actually competitive for a job at a good research lab. As far as I know, these labs typically hire primarily at postdoc level and beyond.