Job Search Marathon

 After a year looking for a Data Science job, my quest has finally come to a successful end and I have some scattered thoughts. One per mile of the job finding marathon. 

Job stats:

1. Applications: 140 (!) Data Science-y jobs applied to (about 50% in MN and 50% remote)

2. Interviews: 54 interviews for 24 positions, I made the final round for interviews for 12 before finally getting an offer

3. Success rate: I withdrew from the final round for 2 jobs after accepting my offer, so I’m going to generously give myself a success rate of ~2/140!

4. Applications per job: Based on what a couple HR people told me, it sounded like there were around 100 applicants per Minnesota job and 300 (!) per remote job 

1. So I guess my success rate wasn’t terrible? 

2. But also it was.

Job-finding process:

5. Given how hard it was for me to get a Data Science job, I can’t imagine how crappy the process is for less “in demand” jobs (especially entry-level-ish). I guess it’s also possible that Data Science isn’t as in demand as people say, but I don’t think that’s it.

6. It was so hard to figure out what was going wrong for me and why I wasn’t getting many interviews or doing well in the interviews. All I knew was it was some combination of:

1. Coronavirus, which meant there were many fewer positions open and the competition for each open position was much higher. 

2. My mediocre (at best) interviewing skills. 

3. Bad luck. 

4. Having less experience than most applicants and only barely meeting most job requirements.

5. Missing some technical skills for some jobs.

6. My resume/cover letter/applications.

7. Everytime I had a referral from someone I knew at the company I got an interview. I’m not sure how I feel about that.

1. On the one hand, referrals make sense to me since it’s hard to really judge a candidate from their resume (or even from interviews). 

2. On the other hand, most of my referrals were from family members and friends that I have never worked with on anything. Even referrals from former co-workers don’t seem that useful, since they are unlikely to not refer someone even if they think their work wasn’t all that good.

8. The hiring process was so varied by company in terms of the number of rounds of interviews they have and what each round looks like. 

1. Some went straight to a final interview after only talking with an HR person for a half hour

2. Others had as many as six different rounds of interviews

3. I had around five interviews with companies who were hiring their first ever data scientist and clearly had noooo idea what they were doing. 

1. It seems like someone should start a DS consulting company that assesses (what a ridiculously looking word) job candidates to make sure they are competent 

4. Of the 12 positions where I got to the final round, there was a lot of variance in technical interviews (I’m not counting technical questions related to my Lewin work, maybe unfair):

1. 3 had no technical interviews

2. 2 had a few technical questions and nothing more

3. 2 had a few technical interviews

4. 1 had a coding/DS online test and then a technical interview

5. 1 had a take-home case study and presentation… but the case study was essentially all about coming up with ideas for how they could use their data… which might be semi-relevant but with a team of 15 data scientists you’d think they’d be more interested in whether someone can code and analyze actual data than what ideas they can come up with

6. 1 had a presentation of any project I had done in the past

7. 2 had take-home data assignments and presentations (as well as technical interviews)

1. Including the one I got my offer at. Coincidence?

9. Despite the variety of interview formats, most don’t seem like they identify best candidates. 

1. The signal-to-noise ratio from a non-technical interview is so low that they really only make sense as a way of weeding out candidates that clearly can’t communicate (which isn’t how they are used). 

2. Even a lot of the technical interviews were focused on knowing random statistical tests or random Python commands. I had one technical interview where every single Python question was about things no data scientist would ever have to know (like dictionaries and lists). 

3. While it’s important to have a good amount of baseline knowledge on coding and statistics/ML, it seems far more important to be able to actually work with and understand data and to be able to learn new things. The take-home data assignments were the only things that came close to accomplishing this. They seemed very close to what you would actually be doing on the job and I can’t really think of a good reason for everyone not to do something like this.

10. I think my lack of success is additional proof the hiring process is flawed. While I don’t have a lot of real-world clear-cut DS experience (mine is more “DS-adjacent”), I’m very confident I can do an above-average job at worst and if a hiring manager could have gotten honest feedback from people I’ve worked with in the past they would have confirmed that I am a quick learner and a strong worker (I realize it’s not realistic to get honest feedback from references)

1. For most employers, the only things their process made visible were my limited experience and not great interview skills. Being able to think clearly about data and figure things out is hard to show, and I never really had a chance in most interviews.

2. For the most part my grad school experience was a disappointment (I could have learned as much as I did in half the time and for free), but I still don’t regret doing it at all because I would have had no shot at even getting interviews without it. Another sign the process sucks and people don’t know how to hire 

11. Why hasn’t hiring improved over time? Especially at big companies you’d think they could afford to invest some serious time and money on improving the process and I would think the returns to any improvement would be so high! Since people rarely get fired nowadays, the costs of a bad hire are so high. And needless to say, the gains from making really good hires are also huge. I’m really curious about this and if it’s just too hard to “improve hiring” or if there’s a lot of low-hanging fruit here.

12. Despite having lots of nice conversations, I have concluded people suck and very few people care about being a decent person (let alone upholding their companies reputations)

1. Of the 10 final interviews I had, 4 of them never got back to me even after I followed up with them (and 2 only got back after I contacted them). This is after having “nice-seeming” conversations with 4-6 people from their company and spending 4-10 hours on the application and interviews (and case studies).

2. The hear-back rate was even worse for the jobs where I had interviews but didn’t make the final round

3. Of the 9 positions where I made the final round but didn’t get an offer, I asked 8 of them for feedback and got one response

4. For one position that sounded really interesting I had a good second round interview where he told me he liked me and that I’d make the final round in a couple weeks. The HR person said she’d get back to me in the next week. After she didn’t, I emailed her and she said she would the next week. Another week went by and again she said she’d get back to me soon after I emailed her. Well that didn’t happen and my assumption is that the position was cancelled. But she never replied to my next follow-up and the hiring manager also never replied so who knows…

5. I realize people are busy, but I just can’t imagine talking to someone for an hour and having them go through a whole process for your position, and then just ignoring them after you decide to go for someone else. The fact that the MAJORITY of people I dealt with did this was pretty depressing and it’s hard for me to come to any conclusion other than that most people are extremely selfish 

13. Despite the above, I would have felt terrible if I got multiple offers and had to turn someone down, since most people were very nice in the interviews and it seems like a major hassle if someone turns down an offer

14. There was a huge variation in how quickly companies moved from my application being submitted to getting to the final interview (I’d say 3-10 weeks). 

1. From an HR perspective it seems like there is a HUGE incentive to get through the process quickly. The longer they take, the higher the chances are a candidate finds another job they end up taking or at least gets another offer they can use to bargain for a higher salary. 

2. For the job I ended up taking, I applied on October 27th and had my first interview there on November 3rd, but then the process wasn’t wrapped up until December 17th. That huge gap meant I interviewed for five positions in December that wouldn’t have been possible had they given an offer to me by the end of November.

3. Anyways, it would be an interesting thing to study if you had the data and it seems like something companies should be thinking about

15. Will it be much easier to get a job next time? I sure hope so. I think it will be somewhat easier, but this process was so long and frustrating that it will probably make me much less eager to go back onto the job market.

Working remotely:

16. For the most part I think remote work should be fine, but I am definitely curious to see how it goes. At my previous job, half or more of the people I worked with were remote or in another office so that makes me feel pretty comfortable from a work perspective

17. The big difference this time is that once everyone’s back in office, I’ll be one of maybe a couple people at the company not working in the office which will be pretty weird, especially at a small company where everyone is working very close to one another

18. It will be really nice to not have to commute to work, both for time saved and not driving in snow. Not great for podcast listening, but once my ankle is better I can get my podcast time on walks instead

19. The biggest downside is definitely that it will be hard to become close with my coworkers and there won’t really be socializing opportunities, both during the workday and outside of it

20. Another downside: being around a bunch of people in an office makes me much more in touch with “normal people” than I otherwise would be 

1. Given my small and very non-representative social (and social media) circle, I don’t have much of a sense for the “general mood” about [Insert Any Topic] and what people are thinking about. Not just on politics, but also basic things like “did you hear about this show people are watching” or “did you hear about this new store/event near us” or “wow D'angelo Russell is so good with his 20 points per game” and other random stuff like that. 

2. I think it’s pretty “important” to have some idea of what people are thinking, but I also think it’s just really interesting to hear people’s random opinions and where they get their information from. Especially in COVID times.

3. My family and extended family is pretty much my only avenue for stuff like this with a remote job (or no job)

Thoughts on working and not working:

21. Unemployment was pretttttttty fun. I never even came close to running out of stuff to do, which I was slightly worried about at first. I spent most of my time reading articles, watching sports, and playing chess. I also did some random online DS classes, random data exploration, and started writing periodically. I think I could have kept all that up for at least another year without getting bored or running out of stuff to read.

22. I rarely played video games during the day which could have easily filled a ton of time with. I also didn’t read books during the day (I only read books at night because that’s what I did when I had a job and I can’t handle change), which is another thing I could do more or less indefinitely. I also spent very little time walking or exercising because of my ankle, which is another thing I “could” have spent more time on. So if I added those three things into my non-working time I think I really could go a very long time without getting bored and wanting a job. I guess I could also watch endless hours of TV shows and YouTube videos as well. So much to do.

23. Now that I have a job how do I fit everything in?!? Over the past year I’ve started reading and doing so many things it’s gonna be pretty hard to decide what to cut. This might be especially hard since I will be working more hours than I did at Lewin and in grad school. I’ve probably played an hour of chess a day for the last couple years which is probably an easy place to cut back. But I’ll still need to cut way back on some combination of watching sports, reading about sports, Twitter, reading non-sports articles, reading The Economist, and reading books. 

24. I am still uncertain whether lack of fulfillment would be an issue for me if I didn’t have a job long term. I know there’s a lot of debate about that around UBI and potential technological unemployment. Maybe fulfillment is only an issue because of societal expectations that people work, which might go away if there really aren’t enough jobs for everyone. I think I am extremely different from most people, so I’m not sure my experience is super relevant to the question of fulfillment would be an issue on a societal level, but it’s still interesting for me to think about on a personal level.

25. I say all that because I think if I really wanted to I could easily work for 10-15 more years and then stop working for the rest of my life. I don’t think I’ll do that, but I do really highly value my leisure time and will definitely consider moving toward part-time work at some point in the not-to-distant future.

26. Even if I’m not super worried about the fulfillment aspect, I still struggle with the ethical side of choosing not to work from both: 

1. the Tyler Cowen/Stubborn Attachments argument that we have an obligation to future generations to grow as quickly as possible (no, I don’t think my contribution to economic growth will be particularly meaningful but they are definitely non-zero) and 

2. the more concrete EA argument that if I choose to work for 20 extra years and donate ~10% of my income every year I can probably save a minimum of 50 lives and improve the lives of hundreds or thousands of people. That’s not even thinking about donating probably all of my savings when I die, which would presumably be a larger sum than the amount I donated during my life. Seems pretty selfish to sit around reading all day instead.

26.2: The end