Which companies have real data scientists?

Visa / Data
yzc

Visa Data

PRE
Visa
yzcmore
Jul 13 60 Comments

I have a MS in DS from top tier CS university, and I have been working as a data scientist for less than 1 year. Daily work includes analytics & AB testing (60%) and building ML models (40%). The current team has some modeling projects, but the larger organization is mainly doing analytics. Their daily work is just SQL+story telling. I would like to transition to a role that is more modeling focused, and has better engineering culture in the next year. Can anyone give me some suggestions on which companies/roles I can apply for if I want to avoid SQL everyday roles? I mean I want to be real data scientist instead of SQL scientist which is in essence business analyst.

Some potential roles are as follows:

Google: Quantitative analyst, Software engineer in highly ML team
FB: DS analytics, DS infrastructure, SWE in highly ML team
LinkedIn: DS, MLE
Amazon: DS, Research scientist, Applied scientist
Apple: DS, MLE
Uber: DS
Lyft: DS
Airbnb: DS analytics/inference/algorithm track
Netflix: DS
Pinterest: DS, MLE
Twitter: DS, MLE
Quora: DS

Any suggestion will be appreciated.

comments

Want to comment? LOG IN or SIGN UP
TOP 60 Comments
  • Credit Karma ckftw
    You applied to a bunch of roles and ended up at VISA? The roles you listed, even the SQL roles, are a huge step up from VISA...

    The most accurate way to gauge what work is actually done is to look at the questions asked in the interview process.
    Jul 13 29
    • Micron ChaChingb
      @ckftw: haha, chops to you confidence buddy. Firstly, Data scientist is a newer hyped field, so the reported data you have or quote to have is a sample vs a population of SWE data. Best of all, show me the data buddy? I don't see links, citations or anything.

      Whatever you are SWE or DS, I bet you love stating opinions without facts.

      If you don't have data, no point in talking. If you do, send it to me, I'll give you a class on analysis, probably you might learn a thing or two?
      Jul 13
    • Credit Karma ckftw
      @chaching lol you just randomly listed a bunch of buzzwords like sample vs population and put them in a sentence acting like you know what you’re doing.

      @ycz just told you that tons of people landed G/FB for software from his class, but 0 for DS. The idea that Google/FB is hiring a bunch of people to do deep learning from random companies because they have experience is laughable. The standards are simply way higher for DS than SWE, because there are 10x more SWE jobs and a ton of majors that pipeline into DS (including CS grads), while very few majors that pipeline into software engineering.

      You’ll see bootcamp grads all over the place at SF tech companies in SWE. The only DS bootcamp grads you’ll see in DS at good companies usually ALSO have a PhD in a quantitative subject.
      Jul 13
    • Micron ChaChingb
      @ckftw: you think sample vs population is a buzz word? Lol. Plus, you've only stated opinions masked as facts. I'm sure you have no idea about Data Science. I would recommend you stop advising people on something you have little to knowledge about. Stick to Software Engineering, SWE person.
      Jul 13
    • Visa / Data
      yzc

      Visa Data

      PRE
      Visa
      yzcmore
      OP
      @ckfw: there is no need to get a CS degree, just leetcode and you will get the SWE job.
      Jul 13
    • Axtria DesiLaunda
      Leetcode skills are not reflected in a resume. How do you get attention?
      Jul 14
  • Amazon snostreblA
    Avoid DS and quantitative researcher at Facebook and Google. Those are essentially SQL monkey/BI analyst roles. Research scientist or machine learning scientist/engineer means you get interesting work.
    Jul 13 1
    • Visa / Data
      yzc

      Visa Data

      PRE
      Visa
      yzcmore
      OP
      What’s the bar for ML engineers at Google and FB? Is there any requirement on year of experiences?
      Jul 13
  • Mu Sigma ERhl51
    I'll give my 2 cents:

    1. Rather than focus on the company try finding out exactly what the data science you apply to does. It is completely possible that you end up doing the same work in a new employer.
    2. Try to find roles in engg/technical teams rather rather than business/ops teams.
    3. Please decide what you mean modelling focused : are you interested in developing models or are you interested in putting models into production (2 are every different beasts). For the former, most companies prefer PhDs with a heavy stats focus. Your CS background might be more suited to deploying models into production. I would also encourage you think about your work in terms of interesting content vs interesting outcomes. Sounds like you are more inclined towards work with interesting content.

    You should think about all this before making a switch.

    PS: learn SQL properly, it's a lot of fun once you master it :)
    Jul 13 4
    • Visa / Data
      yzc

      Visa Data

      PRE
      Visa
      yzcmore
      OP
      Thanks for the insights. Interesting content vs outcome is a good point! Btw I already master SQL well, it’s just I don’t like writing it all day. There is no algorithmic thinking in it.
      Jul 13
    • Mu Sigma ERhl51
      Haha. It can get quite algorithmic as the size of the ADS increases and the data is not well curated (which might be rare in VISA).

      A piece of unsolicited advice, sounds like you don't enjoy story telling because you aren't good at it. Practice story telling better, writing SQL will be more fun and you'll enjoy your current job a lot more.

      Try to become a full stack SWE who can do modelling, IMHO they will command the highest market rate.
      Jul 13
    • Visa / Data
      yzc

      Visa Data

      PRE
      Visa
      yzcmore
      OP
      Actually I like the story telling part of DS a lot, that is even why I go for DS field. I just don’t like writing SQL all day long, it’s not an intelligent work from my perspective. Also the pay is lower for SQL DS.
      Jul 13
    • Mu Sigma ERhl51
      Fair enough. Agree with the low pay about SQL DS. Hence my reco to sort of become a full swe with modelling capabilities. Will give you flexibility do switch between development/data science as you grow in your career.

      Just be aware things aren't rosy on the other side (pure modelling/ML teams). Like half then times you are just sitting around twiddling your thumbs while you wait for data engg teams to build an ADS and when you finally build a model business has lost interest or wants the model to be dumbed down so they can understand it. It's a different kind of frustration when your work has 0 impact on business.This becomes especially critical if you want to grow into leadership roles in data science/analytics.

      Again my experience is mostly based on working with non bay-area F50 companies, might be different in big companies the bay area or start ups.

      Wish you all the best in your job search.
      Jul 13
  • Chegg snt124
    1) just read job description. Most ml roles will emphasize expertise in ml, while most sql monkey roles might say like “knowledge in ml” or not mention ml at all

    2) any ds role with “analytics” in the title is likely non-ml heavy

    3) if a company has roles like research scientist, machine learning engineer, etc., then likely the “data scientist” title is more analytics and those other roles are ml heavy
    Jul 13 4
    • Micron ChaChingb
      Good eye. But quite a lot of companies use generic job description when opening a job role. From a company's perspective using generic role would,
      1. Save time - HRs don't bother Managers and work with what they have. This becomes a lot easier if you are FANG. Reputation brings resumes through the door, hardly the JD.
      2. Projects are not well defined, budget closes soon. Generic description works and once the candidate is in the door we can decide what they do, could be ML research or BA type role.
      Jul 13
    • Chegg snt124
      Yeah I mean there’s no silver bullet answer; these things will help. And ultimately conversations with the hiring manager will provide the most clarity
      Jul 13
    • Micron ChaChingb
      Agreed.
      Jul 13
    • Visa / Data
      yzc

      Visa Data

      PRE
      Visa
      yzcmore
      OP
      Good insights.
      Jul 13
  • Quora hpsB38
    Quora DS, MLE
    Jul 13 2
    • Visa / Data
      yzc

      Visa Data

      PRE
      Visa
      yzcmore
      OP
      Thanks. Is the DS in Quora more focused on AB testing?
      Jul 13
    • Quora hpsB38
      Yes
      Jul 17
  • LinkedIn ex-fb
    Look at the pay and you’ll know what are ML engineer roles and what are BA roles
    Jul 13 1
    • Visa / Data
      yzc

      Visa Data

      PRE
      Visa
      yzcmore
      OP
      For ML engineers, what percentage of time is spent on modeling and general engineering work?
      Jul 13
  • LinkedIn fisheixnd
    With title inflation being what it is, draw your own conclusions on what data science really is.
    Jul 13 0
  • LinkedIn / Eng chiknCurry
    Look for Applied/Research Scientist & MLE roles, NOT DS
    Jul 14 1
    • LinkedIn / Data mSNI75
      Also, not DS - Analytics.
      Jul 15
  • Uber john_list
    The reason it’s hard to get DS job with minimal experience is you really aren’t going to be a good data scientist if all you can do is code or have academic knowledge.

    That’s why they ask so many metrics and product/business sense questions in the interviews. These “SQL data scientists” roles you guys refer to are a closer to a PM in skill set than say to a software engineer. Yes, you need to be able to write SQL and some python. No you don’t have to be an all star coder. It’s a lot more about solving business problems.

    Hire someone out of academia and give them a business problem and enjoy watching them produce a rigorously engineered model, of which will have exactly zero impact on solving the actual problem.
    Jul 14 0
  • PayPal thtP05
    Checkout the PayPal GRDS team they only hire Ms and PHD's. Work is 100% modelling for some teams
    Jul 13 0
  • Micron ChaChingb
    I feel DS is a very generic one size fits all role. If you're interested in ML/DL type of roles search for Applied Scientist/Deep Learning Engineer/Research Scientist from top tier firms. Company product portfolio is very important, FB/Amazon Applied Scientist/Research Scientist roles are heavy into DL modelling whereas their DS roles are BA+simple ML. Nvidia does a lot of ML/DL with Vision work with AS/RS roles.
    Jul 13 0
  • LinkedIn HjT21k
    I say it's more about the team and the work than the title.
    Jul 14 2
    • Visa / Data
      yzc

      Visa Data

      PRE
      Visa
      yzcmore
      OP
      Can you share which teams in LinkedIn has ML work?
      Jul 14
    • Apple Justdance
      Will DS analytics in LinkedIn has opportunity of working in modeling?
      Aug 6
  • New dBBO43
    Openai, Google brain
    Jul 13 1
  • Zillow Group qwiu87
    Zillow’s Applied Science team is heavily focused on algorithms and modeling. They then work with the ML teams for the production side of ML. You get a little bit of everything there.
    Jul 18 0
  • Axtria DesiLaunda
    Following
    Jul 13 0