PAH! Puleeeze…
ML/AI is dominated by 2 companies and their respective teams are located in mtv / mpk. This is an area where a tiny group of people is making 90% of the innovation. The rest of them ain’t really helping out much. 
2 Likes
Making the most noise 
in park and mountain but most of the ML guys are in Seattle.
Companies exodus SFBA or encouraging workers to move out of SFBA is not new and it has been happening since 1999 - dot.com burst onwards. The biggest outflow happened on two downturns.
They move out as they are unable to withstand the high cost of living, be it company or individual. Those who are able to manage/earn more, stay here.
In short, higher earners stay here, lower earners are leaving out to match their income-expense ratio.
But, SFBA is skyrocketing like AAPL stock since then…
I thought it was understood that the #1 innovator in this space is DeepMind in London (yes, they are owned by google but they are treated like an independent company). Also, I think amazon is really really good at deploying this tech internally, they just don’t talk about it.
Out of our sample of more than 36,500 LinkedIn profiles, nearly half (15,747) were based in the United States. A notable proportion of these experts — about one in five — currently work or have previously worked for Microsoft (1,077), IBM (667), Google (697), Amazon (511) and/or Apple (393). Eighty-seven percent of these profiles have at least six years of experience, and almost all of them (97%) say they have at least three years of experience. About a third (36%) of the U.S.-based experts in this talent pool are working in the San Francisco Bay Area.
This report only counts number of people, not their quality. Among American regions it only singles out the Bay Area. To claim that Seattle is the leader, are we supposed to believe more than 1/3 of AI/ML people live in Seattle? Isn’t that a bit far-fetched?
No longer true. You’re extrapolating your experience. Verify with your network.
Regional variation in Salaries.
The regional variations that we see in the salary are only for middle level to lower level jobs. The VP and higher level jobs generally do not have much salary difference across regions for same industry. Between Director and VP, the variation is minimal. Yes, you will see variation across industry for same grade/level positions. The lower the position, higher the variation in salary.
Have you heard of VP and higher jobs going to china to save money? Companies like Broadcom moved their headquarter to Singapore for tax reasons. Have they came back to US after tax cut?
1 Like
I have a feeling the impact of AI/ML is overstated. Similar things were said about Biotech about 20 years ago. Internet and the derivative industry came from a revolution that happens once in 200 years. Hard to replicate the same every 10 years.
Are you in HR? Btw, are you responding to me or @Jil? Size of employees become so large that Director+ need to be with the people. They are not moving to save $$$.
Almost every function and industry employ AI/ML. So can find them all over the places, not limited to SV or Seattle or Austin or Boston. Every (not just American) college are rushing to produce cool AI/ML CS majors.
In one of my previous job, the director lived in Austin and air commuted to work three days in Bay Area.
1 Like
Math grad with good programming skill can do 90% of AI/ML jobs. Not a big deal. The subject upon which it is based (linear algebra and Matrix ) is hundreds of years old. AI/ML will still not replace Creative and Non Repetitive work. So, it will only end up automatic clerical jobs where simple choices can be made based on past experience. Something that happens even today though it is not given a fancy name called Machine Learning.
1 Like
I automate stuff all the time without ML. Anything that requires consistently applying the same logic can be automated. That’s literally a huge percent of corporate work. I’ve even used simple linear regression to determine which inputs have the highest correlation and build prediction models.
The bigger issue with adoption is incentives aren’t aligned. Individuals usually get promoted by managing bigger and bigger teams. Automation allows downsizing of teams. That’s why automation gets lip service, but management doesn’t push for it too much. Management only wants it when a process is so clearly unscalable that they’d never get approval to hire headcount for it.
1 Like
And a bigger budget. And a more complex solution. Is a well known secret to promotion.
Career stagnation.
Another common practice is, best illustrated with example, say you’re a senior manager managing a portfolio, if you screw up, they would either get an outsider Director or an insider manager (to be promoted if successful) to manage the portfolio.
Above type of practices are common in matured companies, is one of the reason why many get disrupted 
by smaller startups.
1 Like
Tell that secret to hiring managers at google and Facebook so they don’t have to waste 300k salary on those people. 
Pure maths guys can’t. Need those with stats minor. ML uses many basic optimization techniques and statistical models. If I’m not wrong, any major can take stats minor.
If the skills are so readily available, do you think companies will pay 300k for a fresh grad? I am not saying AI or ML is super hard. I don’t think so. But to wave it off as “no big deal” doesn’t sound right either. People need to work hard to acquire those skills.
Currently, many statistical implementations are hyped as ML, and many problems that can be solved by traditional structural problem solving techniques are done using such statistical implementations. So ML got a bad name. Implementing an existing ML algorithm is not that difficult, the hard jobs are those that need to design new ML algorithms. Believe latter are those that are earning big bucks.
ML got a bad name where? Last time I check every company says they are implementing ai and investors are jazzed about ai.
Also theory is not everything. How to apply a general ML algorithm in a domain, how to tune it, how to optimize it etc are not trivial problems. Let’s say you work at stitch fix and you want to apply ML to make suggestions to customers. What features do you train your ML? How much weight you put on each? What do you look for to see if your model is good or crap?
It’s like saying I know python therefore all programming jobs are trivial. No.
Allow me to correct my statements.
Bad names to SWEs. Investors are fully conned.