Do ML engineers or sde,ML at FAANG and unicorns build ML models? Or just put models to production and build infrastructure?
Job descriptions differ a lot for these roles across companies. For eg, data science roles at Amazon emphasize a lot on SQL, FB emphasis a lot on A/B testing etc whereas the same does not apply to other companies like say Uber. At Walmart Labs, a data scientist develops models and tunes them. So I would recommend you go with the job description and not the name of the role.
Yeah but different roles have different interview process like : fb, google have mostly leetcode (sde, work on ML) .Uber,Apple have interviews focused towards ML (data scientists work on ML). So for similar work interview processes differ which makes it difficult to prepare.
IMO anyone who can do MLE can easily pass leetcode, not so much vice versaz
It depends on the team and company