I got a DS job at Lyft. It seems that they changed the title recently so it’s more like a statistics heavy product analyst role. Personally, I’m also interested in being a machine learning engineer, i.e. model building and deployment. I have the skills of SQL, Statistics and Python, and have done some POC work at my current job implementing ML models. How hard is it to switch from DS to ML Engineer? Will the skill sets vary significantly? I’m still exploring my career path and wanted to get some input from Blinders. Thanks!
SJW is right. Most DS are glorified analysts who don't have a proper grasp of the mathematics and theory behind the 'import from' modules they recklessly use, if they even have a decent grasp of a production grade scripting language (not R), that is. I've seen that any analyst who makes enough of a fuss and hoopla about how they're unfairly classified as an analyst and "deserve" to be DS eventually get the title and associated pay bump. Not just here but at my last company too (also FANG). 7/10 of these guys and gals don't know the difference between frequentist and bayesian mindsets, cannot fully explain statistical Power, nor can they pass calculus or linear Algebra 101. What they are good at is being good business analysts: thinking of improvements and metrics to measure them, and SQL.
Well.. I see that at workplaces too and it could be frustrating. And that’s what I have been trying to avoid. Not all DS are like that. Hope you don’t judge us by the title.
Not the case at FB. The two teams that I have been on have been over 50% math/physics PhDs. They're admittedly not stats experts, but have a demonstrated ability to pick up complex math with little effort.
ML Engineer is the hot thing right now, but I don’t think DS will hurt your chance of getting an interview. In all likelihood, the company has a good process (or Leetcode) in place to test your CS, so your job helps in speaking to the other part of the role. I would not stay much more than 1 year though.
Googled it. That’s a long way to go but I’ll take baby steps. Thanks.
At GE/Data, how's the core DS team at Lyft (the original DS team)? Do you know what the interview process for those roles is like? Also, what's your background like?
How was the interview at Lyft?
ML is fast moving away from DS to Eng, given that DS only really know how to prototype and the real value is having things actually delivered. As SWE ML you’re still expected to know ML pretty well, but also Eng to at least the same degree. DS is essentially becoming associated with analytics and not so much ML any more.
Thanks for the reply. I guess if I decide to switch later I’ll need to start with some algorithms and side projects. What worries me is that when recruiters see DS title I won’t even get a chance for the next steps.