Close

Complete Captcha

• 11 months - 253 posts - Latest - RSS
• Thread: 624 Goods vs 234 No Goods

1. Economist
   2b57
   

   I replied to a thread about leaving academia for industry (https://www.econjobrumors.com/topic/failing-to-leave-academia? -- I'm f3b4). I was an econ AP with a low-prestige PhD at a failing school and I couldn't publish my way out.

   Seems like lots of peeps are interested in the transition. When I was gearing up to go I got lots of ideas from this message board. Hoping I can give back with this AMA. Also hoping I can get ideas from other bros/broettes who made the transition. I'll do my best to answer all questions in as much detail as I can without giving myself away.

   11 months ago # QUOTE 21 Volod 1 Vlad !

2. Economist
   bc5f
   

   What's a typical day look like? When do you arrive/leave work? Do you have time to brush up on skills?

   11 months ago # QUOTE 1 Volod 0 Vlad !

3. Economist
   00fd
   

   What was your X-salary multiplier vs being an AP? I was a UK AP, and hated life there so I left for industry in another country. I don't regret it and would never go back. It is miserable there. Even more so in academia.

   Was nice to get the 3x salary boost too! I enjoy life now.

   11 months ago # QUOTE 2 Volod 0 Vlad !

4. Economist
   2b57
   

   What's a typical day look like? When do you arrive/leave work? Do you have time to brush up on skills?

   I'm remote. Roll out of bed right before morning meetings. I have to carve out time to learn new stuff. To get motivated I look at new job postings and talk to other industry friends (most of whom I cold-wrote on LinkedIn) to figure out what and how much to learn.

   Typical day: some meetings that I just work through or browse internet, write code for ongoing projects, maybe skim a paper. Trying to spend more time learning my firm's business. Domain expertise outweighs technical expertise pretty quickly.

   11 months ago # QUOTE 10 Volod 0 Vlad !

5. Economist
   2b57
   

   What was your X-salary multiplier vs being an AP? I was a UK AP, and hated life there so I left for industry in another country. I don't regret it and would never go back. It is miserable there. Even more so in academia.
   Was nice to get the 3x salary boost too! I enjoy life now.

   Made 80k as an AP. Now make 180k (salary + bonus; firm is private, no equity) with yearly raises. Three weeks paid vacation, then some sick and personal days. Most days I can get away with doing just about 4-5 hours of real work. If I need an extra day off I just ask my boss. I have to be in good graces to get it.

   11 months ago # QUOTE 5 Volod 0 Vlad !

6. Economist
   f585
   

   What did you fill your portfolio with? Empirical econ projects or cookie-cutter ML exercises?

   11 months ago # QUOTE 1 Volod 0 Vlad !

7. Economist
   2b57
   

   What did you fill your portfolio with? Empirical econ projects or cookie-cutter ML exercises?

   Projects. Put the replication codes for my papers on GitHub, alongside some side projects. I billed myself as an economist with strong coding skills. Hard sell if you have no pubs but you could write blog posts or something. I knew I couldn't compete with a CS/stats candidate so I didn't bother drawing focus on my ML skills (which are adequate, not extraordinary). Cookie-cutter ML stuff is useful for your own education if you have no ML experience but as a signaling device they are a waste of time. Anybody with half a brain will wonder if you just copy-pasted somebody else's Kaggle submission.

   Lucky for us it's a good time to be econometrician. Friends at other firms tell me they recruit data scientists who are strong in causality, experiment design, etc. Especially true for jobs that require PhDs. Incredible how far a solid understanding of linear models, confounding and causality will take you.

   11 months ago # QUOTE 13 Volod 0 Vlad !

8. Economist
   d6aa
   

   Do you ever regret it? Not necessarily asking whether you think it was a bad choice sitting here today (I take it you are happy with the choice overall), but to what extent are there moments where you wish you were back in academia?

   11 months ago # QUOTE 1 Volod 0 Vlad !

9. Economist
   3252
   

   ljl at "causality".

   11 months ago # QUOTE 2 Volod 5 Vlad !

10. Economist
    3252
    

    seems like data science is shifting away from ML stuff and moving towards more econometric style techniques relying more on business knowledge than just shoving data down a black box. I guess this is good for economists.

    11 months ago # QUOTE 1 Volod 6 Vlad !

11. Economist
    0d7b
    

    What did you fill your portfolio with? Empirical econ projects or cookie-cutter ML exercises?

    Projects. Put the replication codes for my papers on GitHub, alongside some side projects. I billed myself as an economist with strong coding skills. Hard sell if you have no pubs but you could write blog posts or something. I knew I couldn't compete with a CS/stats candidate so I didn't bother drawing focus on my ML skills (which are adequate, not extraordinary). Cookie-cutter ML stuff is useful for your own education if you have no ML experience but as a signaling device they are a waste of time. Anybody with half a brain will wonder if you just copy-pasted somebody else's Kaggle submission.
    Lucky for us it's a good time to be econometrician. Friends at other firms tell me they recruit data scientists who are strong in causality, experiment design, etc. Especially true for jobs that require PhDs. Incredible how far a solid understanding of linear models, confounding and causality will take you.

    did you do any Coursera classes? Udemy? cloud certifications? which ones would you recommend?

    11 months ago # QUOTE 0 Volod 0 Vlad !

12. Economist
    2b57
    

    Do you ever regret it? Not necessarily asking whether you think it was a bad choice sitting here today (I take it you are happy with the choice overall), but to what extent are there moments where you wish you were back in academia?

    No regrets. I was simply not a great researcher. I was a good teacher but the market doesn't value teaching. Too bad. I only knew that if I stayed at my school I would never have money, professional respect, interesting colleagues, or any of the other stuff we need to feel good about ourselves. I left academia because I felt I had no choice. So it goes.

    Would I take an R1 job over industry? Absolutely. Tastes vary, but for me, a good academic job is way better than an industry job. Tenured life at a good school seems awesome.

    Here in industry I worry a lot about staying close to the state-of-the-art. I have no idea what my career path will be. Sometimes I feel the walls closing in. I have to remind myself that most people live this way. Life in academia gave me a false sense of the world.

    11 months ago # QUOTE 14 Volod 0 Vlad !

13. Economist
    2b57
    

    ljl at "causality".

    Laugh all you want. Then browse DS postings on LinkedIn at major firms. Causality is front and center.

    11 months ago # QUOTE 4 Volod 1 Vlad !

14. Economist
    5f8b
    

    seems like data science is shifting away from ML stuff and moving towards more econometric style techniques relying more on business knowledge than just shoving data down a black box. I guess this is good for economists.

    And for getting actual business results.

    11 months ago # QUOTE 0 Volod 0 Vlad !

15. Economist
    6877
    

    Do you still keep up with academic research, even if it is just casual reading?

    11 months ago # QUOTE 0 Volod 0 Vlad !

16. Economist
    2b57
    

    What did you fill your portfolio with? Empirical econ projects or cookie-cutter ML exercises?

    Projects. Put the replication codes for my papers on GitHub, alongside some side projects. I billed myself as an economist with strong coding skills. Hard sell if you have no pubs but you could write blog posts or something. I knew I couldn't compete with a CS/stats candidate so I didn't bother drawing focus on my ML skills (which are adequate, not extraordinary). Cookie-cutter ML stuff is useful for your own education if you have no ML experience but as a signaling device they are a waste of time. Anybody with half a brain will wonder if you just copy-pasted somebody else's Kaggle submission.
    Lucky for us it's a good time to be econometrician. Friends at other firms tell me they recruit data scientists who are strong in causality, experiment design, etc. Especially true for jobs that require PhDs. Incredible how far a solid understanding of linear models, confounding and causality will take you.

    did you do any Coursera classes? Udemy? cloud certifications? which ones would you recommend?

    None of the above. I needed my PhD to be my signal. How could I claim my PhD gave me valuable skills if I needed MOOC certificates? Although I did take a DataCamp course in SQL for my own education. Now that you mention it, I could see how certifications in stuff like Azure/Hadoop/whatever could be useful. My sense is that hiring managers who know their stuff prefer to see nice GitHub repos of your own making rather than certificates.

    I learned stuff by working on projects on my own or replicating examples from econ textbooks in R/Python. It helped that I enjoyed programming. I had no background in it before my PhD.

    11 months ago # QUOTE 1 Volod 0 Vlad !

17. Economist
    2b57
    

    Do you still keep up with academic research, even if it is just casual reading?

    Yes. This is critical. You have to live up to people's expectations of a PhD. It's a research degree. So you better know how research is evolving. That said, I read less broadly than I used to. As an AP I would browse journals just for kicks. Now that I have more pressure on my time I have to be more strategic.

    11 months ago # QUOTE 2 Volod 0 Vlad !

18. Economist
    2b57
    

    seems like data science is shifting away from ML stuff and moving towards more econometric style techniques relying more on business knowledge than just shoving data down a black box. I guess this is good for economists.

    And for getting actual business results.

    Indeed. Data science is just an input to a manager's decision-making function. They want to know if sales will go up if they press this marketing button. Your job is to say yes or no.

    11 months ago # QUOTE 2 Volod 0 Vlad !

19. Economist
    2bcd
    

    OP is a good dude

    - DS bro who left with MA

    11 months ago # QUOTE 8 Volod 0 Vlad !

20. Economist
    0acf
    

    What did you fill your portfolio with? Empirical econ projects or cookie-cutter ML exercises?

    Projects. Put the replication codes for my papers on GitHub, alongside some side projects. I billed myself as an economist with strong coding skills. Hard sell if you have no pubs but you could write blog posts or something. I knew I couldn't compete with a CS/stats candidate so I didn't bother drawing focus on my ML skills (which are adequate, not extraordinary). Cookie-cutter ML stuff is useful for your own education if you have no ML experience but as a signaling device they are a waste of time. Anybody with half a brain will wonder if you just copy-pasted somebody else's Kaggle submission.
    Lucky for us it's a good time to be econometrician. Friends at other firms tell me they recruit data scientists who are strong in causality, experiment design, etc. Especially true for jobs that require PhDs. Incredible how far a solid understanding of linear models, confounding and causality will take you.

    What are a me forms/jobs that recruit for experimental design and causality?

    11 months ago # QUOTE 0 Volod 0 Vlad !

Reply »