I am curious why you're transitioning off from Datadog?
A monitoring solution based on Grafana and Prometheus tends to have a high initial setup cost plus on-going maintenance since it's self-hosted and lacks some of the features that Datadog offers out of the box (alert customization for instance).
Custom metrics on Datadog are the main reason. We are a game company so we have all the metrics from the client games going through our server, which then send them to Datadog as custom metrics. This lets us monitor our client games in real time. When we started with Datadog many years ago it was fine, but then about 1.5 years ago they changed their billing to only allow 100 custom metrics per host. We were using about 2000 custom metrics per host. We had to aggregate the custom metrics down to an almost non-useful level to stay within the 100 custom metrics per host.
Also the $15 per host starts to add up if you want to run many smaller hosts rather than a handful of large ones.
I've been on many calls with my Datadog account rep and they refuse to budge on this. I would stay with them in a heartbeat if it was reasonably priced for what we want to do.
I have both worked with and interviewed people in their 40s when I worked at a FAANG. Despite what's advertised by HR, culture tends to be org/team-dependent at bigger tech companies. I believe that you should be more concerned about culture fit which is something you can assess during the interview. If you decide to interview for an individual contributor position, I recommend that you brush up on algorithms. You can either target a company on LeetCode or go through a generic list of problems (here is the one I use https://theinterviewlist.com)
In addition to the resources others have suggested here I recommend going through this list of problems that I put together for the coding portion of FAANG interviews: https://theinterviewlist.com
Looking at glassdoor.com and my personal experience, salaries are much lower than what people are claiming on HN
A major problem with Glassdoor salaries is the absence of a date range. The salaries they show is the average over the period of time starting when they started collecting salaries. So yes the salaries on Glassdoor tend to be lower than the current average.
A lot (too many) people on HN are claiming $200k base salary per year as being totally normal.
I agree. It is pretty high. My own reference for base salaries is my own website where I collect H-1B salaries. We also have an option to only show the salary distribution starting on a given year. For instance, for a software engineer at FB, the average base salary is $149k if you only include salaries reported after 2015 [1]. At Netflix however, Sr. Software Engineers make $200k on average[2].
Consider also that most engineers are new grads. Glassdoor and friends should display an "average" that's very slightly higher than the new grad base salary, since most practitioners entered the field recently.
$200k base salary could well be totally normal for people with more than 10 years of experience, who are barely represented on Glassdoor.
Do you know when the paperwork is filed? ie are the listed H1b salaries based on first day of work at the company, or pay after multiple years? I'm trying to figure out if this is the money the company used to get the talent or if its reflective of their own internal raises as well?
The salaries shown on H1BPay are the salaries that the company promises to pay or is currently paying the H-1B applicant.
There are 3 common cases of when the paperwork is filed. First, foreign students who graduated from American universities in certain fields can work for a certain period of time after which they can transition to H-1B. So the salary is pretty much what they are paid when the application was filed which is somewhere between 1 and 27 months. Second, people hired from overseas or within the US who aren't new graduates and don't have a current H-1B visa. Their salary would be what the company offered to get the talent. The third case is when you transfer you H-1B visa where an application is not required so there is no change to the salary in the application.
Fantastic. The rate of internal raises over the years at a company is something I've not seen a lot of statistics on. Your responses was exactly the claritrty I was looking for! Thanks!
There are relatively easy and cost effective ways to filter candidates from the supposedly less-than-stellar schools. Personal projects, open source projects contributions and programming contests are a good way to easily gauge candidates for a software engineering position.
I once encounter someone on reddit who says he purposefully ignore personal projects and open source projects contributions when making hiring decisions, because he thinks it is a white male privilege.
That seems like a cartoonish summary, or maybe they're just a cartoonish person. But I think there's some truth in there.
It's undeniable that open source is mainly white dudes. There are plenty of women who will tell you that they don't feel welcome in open source, and so don't participate or have stopped participating. So if you use open source participation as a hiring filter, it will have a bias.
Personal projects are a sign of having time and money. That is also correlated with being a single white guy. Parents, and particularly single parents mostly just laugh when you talk about free time. Making humans from scratch is their personal project.
That's not to say that these aren't good things to look for. I look for them myself. But if you use them as a filter (as opposed to one of many positive signals), then you definitely are biasing your hiring to middle- and upper-class white males.
I object to the racism and sexism in the sentiment, but it is a sort of privilege to have access to computers and time to create personal projects on them. But, it's still a good predictor for job success, so unless it results in discrimination against protected groups, it should still matter for hiring.
On the flip side, even if personal projects don't matter for hiring, everyone interested in the industry should be pursuing personal projects, even if it takes buying garbage computers (as I did as a kid) and repairing them to get a learning platform. Personal projects and open source are more accessible than people might think.
It is surprising that you thought of this as a "cool AI job offer". I have two remarks here. First, the email sent to the class is barely an invitation to send resumes. Something many programmers/CS Students with online presence experience on a regular basis. Probably not from a Stanford Professor but at least from major companies recruiters. It would be interesting to know how many will actually make it through the screening, phone/on-site interviews and get a job offer.
Second, I registered for the Machine Learning course (I am not sure if the same applies to the AI course) and I compared it with the actual ML course at Stanford (CS229) (I mainly looked at Youtube videos of Andrew[1] as well as Assignments/Midterm[2]). The latter is by far more advanced and theoretical. The assignments tend to test more than basic comprehension of the material presented in the lectures, which is exactly what the online course reviews tend to evaluate. They require strong mathematical knowledge and obviously a minimum level of creativity/intelligence.
"It is surprising that you thought of this as a "cool AI job offer"."
I don't. That part of the post was written with tongue firmly attached to cheek. If that tone didn't come through, that means I have to improve my writing.
The online ML course is CS 229A (which is also an actual course at Stanford. The online version is close to the Stanford course).
The "tough" version is CS 229 (no 'A' at the end). I registered for the ML course thinking it was an online version of CS 229 and dropped out when it was confirmed to be 229A. In my politically incorrect opinion, 229A is close to worthless. The math is important in real world ML. This course included gems such as "if you don't know what a derivative is, that is fine".
The online AI course is almost exactly the same course as Stanford (CS 221), minus, of course, the programming assignments. It is an introductory, broad based course, and it does the job well (imo)
The online DB course is almost (if not exactly) the same as Stanford CS 145. I think this was the best course of the three.
All courses track the corresponding Stanford courses.
> 229A is close to worthless
> This course included gems such as "if you don't know what a derivative is, that is fine".
It also included other gems like debugging models with learning curves, stochastic gradient descent, artificial data and ceiling analysis. I have not come across practical things like these in more mathematically oriented ML books that I have tried reading in the past.
Interestingly, your arrogance is in sharp contrast with the humility of the professor, where he admits in places that he went around using tools for a long time(like SVM) without fully understanding the mathematical details.
I'd hardly call it worthless myself. It lacks a deeper analysis of all the methods that are used, but using them can sometimes be a greater challenge.
I did the AI course and the ML course and find it a great way of getting a little overview of the subjects, so when I study on my own, I have a little direction.
On the other hand, it you already know what a derivative is, you already went through all the lineal algebra stuff, have an idea of numerical methods, etc, I appreciate not wading into those side areas. Specially if you have kids, a dayjob and doing the AI-class at the same time :D
> "if you don't know what a derivative is, that is fine".
A bit of me died when I heard prof. Ng say that. However, I had committed to finishing ml-class and I did. As of now, I'm glad I went through with it. I felt like I was learning all these cool AI techniques that I hadn't heard about. However, the proof is in the pudding. The question is will I be able to take a real world problem and apply what I learned in that class to come up with something interesting? If I can't you are probably right. My perfect record would only be worth the paper it's printed on and the money I paid for the course!
I'm not pointing fingers at Prof. Ng. or anyone here. It was an experiment for Stanford and an experiment for me. I know I am looking forward to the courses next year :).
