Meta reviews

* Your job satisfaction (and your salary) can change greatly depending on where you are in the company. My technical title was substituted with a non-technical "specialist" equivalent that paid less, resulting in a 20% bump in my salary when left, despite downgrading from a "lead" level position at Facebook to a "senior" level position at my current job. * Management is often promoted based on seniority rather than competence. * "Most fast" mentality leads to duplicated work, lack of documentation, and general confusion. Senior ICs are expected to spend much of their time project managing to make up for the failure in organization. I often found during requirements gathering for a project that part or all of it had already been done elsewhere and everyone had simply forgotten about it. * The culture of open communication makes it seem like employee feedback is valued and encouraged, but actions are rarely taken on that feedback. The above issues weren't as bad when I first joined, but my last year there was a nightmare filled with so much toxic stress and bad politics that I regularly broke down crying when I got home from work. On the other hand, I have friends elsewhere at Facebook who do the same job and love every minute of it. It all comes down who your managers and directors are. Personally, you couldn't pay me to go back.

If you are a data scientist, you need to really care about two things: - Impact. Unlike SWE, who can prove their impacts by finishing pre-planned coding projects and rolling out pre-planned features, data scientist does not automatically get credit if you only finish beautiful analyses. You have to suggest eng team to apply your suggestion from analysis, and prove how much of the product growth are from your analysis. So sometimes your impact can be subjective. - Skillset. You may be asked to do a lot of ad-hoc analyses from eng team, which I am sure are very important and can guide engineers whether to implement certain features, but that will prevent you from doing some deep dive analyses and learning some new techniques in data analysis. Maybe after a year or two, you will find that you did not build any statistical models or haven't touched any machine learning. But that's the skillset what other companies will ask for when you apply new job. Who cares you did how many ad-hoc analyses or AB test or wrote how many data pipelines. It's the fancy machine learning model that will earn you respect. Also, total pay is lower than SWE for the same level.