Job Market in the USA through my eyes
Hello, community! In my free time, I decided to share with you the story of my job search in the USA in 2023-2024. Currently, I live in Berkeley, California. I have been here since the summer of 2021, and this is, so to speak, my second job search.
I currently work in RealmI am engaged in the creation of features around generative AI and their introduction into production. Two years of fairly intensive work in a real Silicon Valley startup. Before that, I worked in Skoltechin the computer vision group of E. Burnaev, with him and other incredibly talented people.
Intro
In this article, I will share my experience of job hunting in the US, from working with recruiting agencies and applying directly to companies, to participating in hackathons and networking at various events. I will share my successes, failures, and lessons learned during this period. I hope my story will be useful to those who are in a similar situation or are simply interested in the American tech job market.
Work is not a wolf
After two years in a startup, it became more than obvious that the longer you stay in this place, the greater the threat of being forgotten as a professional. Where the wind blows, there the startup moves. This approach has the risk of a huge blowout in case of success: employees-engineers can end up in management positions in a large company or simply with good money. But this is not always the case, and professional growth, in my opinion, is a much more predictable thing. I preferred to find a more populous company with a clearer product and sales processes.
Two sides of the same process
Now directly about the job search. I have heard from fellow students and former colleagues about how fiercely employers from Russia hire, which generally extends to the near abroad. In America, everything is completely different now, or at least it is not so noticeable.
I personally receive a lot of job inquiries. Which is absolutely irrelevant to this article, but the fact itself is a nice bonus.
Recruitment agencies
Recruiting in America is very widespread, but I can't say it is developed. Here I will describe the process of working with professionals who find you themselves with the aim of receiving a commission in case of your employment. Distinctive features of recruiting in the USA that are immediately obvious:
The more senior the recruiter, the more likely he is to be a man, and in general men predominate in this field.
Recruiters are willing to spend an hour on the phone with you to discuss all the positions they have and how you might be a good fit, but in the end they may never say yes or no to whether we'll work together again.
A Chicago employee might be trying to fill a position in San Francisco, but if you're in San Francisco and want to get a job in nearby Los Angeles, that's a problem.
In general, despite the above-mentioned problems, the chance of getting an interview through a recruiter from an agency is much higher than when applying independently. In addition to the fact that with an external recruiter I got at least 5 companies for interviews, with them you can simply practice self-presentation, which is objectively better than a dry refusal from an automated system.
We don't take it to heart. Things are like this because of the high competition
Here is an approximate timeline of working with the agency:
January 30 – connection with a representative on LinkedIn, exchange of pleasantries,
On April 15, he offers two vacancies, which we are working on in sequence,
We didn't get along with either company and the interviews stopped somewhere in the middle of the process. But it was a good training session.
I will tell you about the second side, about working with internal recruiters, below.
Direct Hire
From experienced professionals and from posts on the Internet I often hear about the approach:
make a table with the top companies with offices in the region,
in the careers section apply for all relevant vacancies,
wait for answers,
track progress.
The approach is good, but I use it only partially. The system and progress tracking are useful. But, in my opinion, there are a dozen companies where I would like to work without any “BUTs”. Companies like NVIDIA, Meta (FAIR), Apple, OpenAI, Anthropic… They are targeted, preferably with referrals, it makes sense to apply. In other cases, what stops you from randomly applying based on keywords?
Let the recruiter do their job and determine how suitable you are!
In a highly competitive environment, a call from a company's internal recruiter in 90% of cases means a 90% match with a vacancy for me.
A small note about competition. It is no secret that people from all over the world apply to the US, nothing prevents a simple online application. On the other hand, here is an example: a third-year business school student at UC Berkeley, a 100% American, says that his business club invites students like him to intern at large consulting companies. But with an interesting condition – the internship will be in the summer of 2025, a year from now! That is, places are limited.
Applying to companies directly, I managed to get interviews only at Meta and NVIDIA. Moreover, those vacancies for which I had referees were rejected. I managed to get interviews for other vacancies later.
NVIDIA
I have been applying to NVIDIA since 2019, back when an office dedicated to development, and not just testing hardware, was just starting up in Moscow. Back then, I applied for a solution architect, an incredibly popular position at NVIDIA, and during the interview, a little embarrassed by such a possible status, I blurted out when asked how to check the state of the GPU that I needed to do “some kind of blood test.” Spoiler: they asked about nvidia-smiwhich I was familiar with, but for some reason did not think of such a simple answer.
Then we had a good laugh with the interviewers over my answer. It is worth noting that such a practice, usuallyis not found in interviews in the US.
NVIDIA was a super interesting experience. I applied for some positions with a referee. Despite the fact that Mr. Huang talks about the fantastic position of the company in the era of AI and the unprecedented (re)valuation of the company at 3 trillion dollars (at the time of the interview in March it was only 2 trillion), only one out of 30-40 responses advanced to the interview stage. The position was supposedly Worldwide, Remote, but in fact they were hiring from the Polish office, and during the interview one interviewer taught another, younger colleague, how to conduct interviews. I can’t say that I didn’t solve the problem, but in the end it was a refusal. Later I managed to agree on a conversation with a recruiter from Santa Clara, where they calmly went through a prepared script about what could go wrong and what I should pay attention to if I ever get a chance to get an interview with them.
Meta
At the end of April, a classmate of mine recommended me to Meta, and I chose three vacancies that I thought were relevant. I was rejected for them, but a couple of weeks later, a recruiter called me with an offer to go through the selection process for the role of a machine learning generalist at level E5. This is where things got very interesting… After 5 years of periodically applying to Facebook, I not only spoke to a recruiter for the first time, but also got an interview!
The first interview is a screening. During it, you solve two algorithmic problems with leetcode. Based on the results of this interview, it is determined whether you will go on to the on-site, which is held off-site via Zoom. The screening is preceded by a mock screening – the “same” interview, but it does not count towards the overall score. Looking ahead, I was given two problems, leetcode medium and leetcode hard. The interviewer is extremely friendly towards you and gives reasonable hints, just like a teacher would do.
Interviewing potential employees with the ability to provide advice, but without any responsibility. Isn't that a dream job?
At this interview, I caught a hard flashback from my first year of MSU CMC – the problem was solved using prefix trees. I used them for the first and last time at that time, there, at the university. Despite this, the trie data structure is called very useful in the Meta preparatory materials. Fortunately, the flashback was very detailed, and I solved this problem. The problem is similar to this
Real screening, of course, looks completely different. This is expressed in the procedure, and in the fact that the tasks are leetcode easy, and in additional questions that require rewriting the algorithm if the candidate has not agreed on all possible data in advance…
The processes are quite drawn out, and this is deliberate. After the screening, you need to book a conversation with the recruiter, where he/she gives you feedback and says that you will work with the next recruiter, with whom you need to book a call, obviously.
Communication with an onsite specialist is an hour-long telephone conversation, the length of which is apparently determined in advance. During it, the candidate is told about each of the 4 interviews that await him, sometimes retelling text material, but with the possibility of interaction. They say that it is imperative to highlight month for preparation. My task was to negotiate at least so that I could pass them in at least 3 weeks.
In general, everything takes a very long time.
Behavioural interview
At Meta, I went to on-sites (which are actually off-sites), and there were two coding interviews, a machine learning system design, and a behavioral interview. I had heard warnings about the complexity of these behavioral interviews from many places before, but I had never experienced it on my own skin, exactly as they do in large companies.
This beast is actually not simple and is difficult to tame. I have looked through a lot video on this topic. It is very difficult to prepare for an interview, where the best advice is considered to be, “don't try to reverse engineer the questioner's questions” and “be genuine”. Firstly, how can you fully answer a question if, speaking about the situation, the questioner knows what, in his opinion, is the right answer? Secondly, how can you be objectively genuine, if even in the workplace we all usually adhere to frameworks and roles, what can we say about communicating with a random stranger?
I have developed the following approach, that it is not necessary to understand the answers to these questions. Behavioral interview is a complicated thing, but as a framework for a capitalist it is useful. People understand each other within the framework of common rules. So let's stick to this.
And about how to pass such interviews: get the most frequently asked questions from the Internet, prepare stories as arguments for the Unified State Exam in Russian. And match these stories to the questions in the most appropriate way
How is the problem of deceptive appearance connected to the fact that Raskolnikov beat the old woman? A little more connected than the problem of the egoism of a man in love.
About the interview in general
Simplified statistics are as follows: I applied both independently and through referrals to about 200 companies (in 8-9 months). I only reached the final stages in three: Meta, Helm.AI, Rock.
By final here we mean that there are no more interviews, in total: here is your refusal or here is your offer. Quite a few companies simply disappear on their own, or I do not complete the test task, which requires 5 hours of work.
The roles in all companies were either senior or leadership. As a rule, I dropped out at the last stage when communicating with management. At some point, it becomes difficult not to solve problems in leetcode during interviews, especially when there is ChatGPT, which explains the solution to any problem in any language, when interviewers, as if by agreement, ask the same problem in DFS with islands, and when they mostly prefer easy levels and in rare cases medium.
Analyzing communication with managers, I feel that it is very important in a competitive environment to have more relevant experience than others. More confirmed cross-functional work, which is very common in large companies. In my case, it is the opposite:
Almost 30 years and almost never worked in a corporation. Is this a virtue or a vice?
From here I have a general conclusion on how to achieve the goal: get a job in a corporation. During high competition for positions there, gain experience in cross-functional work, perhaps in another place. And continue applying.
Networking
Obviously an important and, most importantly, versatile topic when looking for a job. Connections are a useful thing. While I was working in Russia, there was a period when a good friend of mine offered me to join his department and work for a higher salary than I had at the time. I liked this possible “plan B”, but, of course, it is better to move your own way, and that is what I did.
This type of networking won't fly in the US, of course, but if that friend is the CEO of a startup that has already raised investment, there's little stopping him from doing this type of hiring.
Returning to the topic, I would like to talk about networking, which implies professional connections with an unspecified purposeThis is exactly how I formulate it for myself when I go to some event with strangers.
Meetups
Let's start with something simple. Meetups are not the best place to build professional connections. Often, people who come to themed meetups are not in the know yet and want to learn about it. For example, at the data science meetup in San Francisco, you can mostly meet very new developers, students, or non-developers/analysts. Of course, there are people there who are interesting from a professional standpoint, but to do that, you need to talk to 20 people for 3 minutes. I did this during my first job search and at least met one developer guy that way, with whom I still keep in touch.
Hackathons
I have been participating in hackathons since 2018, when Mail.RU there was a school of data analysis “Technosphere” at Moscow State University, and when they held events like SmartMailHack. It was a great place to work together with people you barely knew and become good friends, understand who can do what and what you can achieve together. At that time, it was primarily a new experience and the drive to compete for prizes.
At this stage of your career, participating in hackathons has different purposes. First, not all hackathons in the Valley offer prizes. Second, not all companies that host them are as actively hiring as they once were. Mail.RU.
In the Bay Area, as a participant, I was able to attend three hackathons in two AGI House. And even at one of them, my team and I took second place. The prize for us was a private dinner with Richard Sochercreator You.Comat his ranch. That's some good networking.
In addition, you get to know the guys better with whom you managed to put together a team.
Here, hackathons are held much more often in private hacker houses than in the offices of Google or Facebook.
As a jury member of hackathons, I am periodically invited to the UC Berkeley hackathon (CalHacks) and the UC Davis hackathon (HackDavis). I go there not for networking, but to see what students are doing and, perhaps, to help somehow.
Conferences
I find this subsection of networking to be the most fruitful in terms of expanding your circle of acquaintances, especially when you get to major events in your area of interest. For me, such conferences were GTC 2024 And Fully Connected 2023It is especially pleasant if they are invited for free or paid for by the employer.
At such events, the most interesting from the point of view of job search are the reports of employees of large companies about their products, which are essentially similar to the answers to a system design interview. Also interesting is an exhibition with stands, where you can meet employees of smaller companies that you have never heard of before, but they are doing something interesting, and immediately take a business card for possible contact regarding work.
I was never invited to a job interview from the conference. But I was invited to other conferences for free, where I met respectable people, high management, like those who made Oculus VR and IBM Newton. These people have assistants who answer messages on LinkedIn for them.
The guy from Oculus VR told the story of how John Carmack (the Chuck Norris of programmers) brought code from his previous employer to Oculus, which ended up costing the company $500 million in court.
What is the motivation behind this?
At the moment I have not received any offers from a large company, but the process continues. I started this business almost 9 months ago, and it seems that gradually something is starting to happen. It's all a matter of time.
I wouldn't be surprised if my text makes it seem to many that the path is irritatingly difficult and perhaps not worth the effort. Why not stay in the current company and calmly collect your salary? There are two arguments against this that are decisive for me:
At the graduation ceremony of the Bachelor's degree in Computational Mathematics and Cybernetics, a professor of the ternary informatics laboratory spoke to us, the graduates. He formulated the answer to the question of what to do next very well: be where there are people from whom you can learn something! And in large companies in California, there are very smart people working, and in addition, they are certainly well-disposed towards you.
Huge salaries at senior levels, which are not offered except in large companies. For example, today E5 at Meta, it's $450,715 per year in total compensation. And not just salary, but the overall quality of life of employees, whose convenience they fight for. On the territory of Meta Campus, for example, you can go to the dentist (not that it's a very adequate bonus, but still), and change the oil in your car (this is a little more relevant). And also, when you're stuck in a traffic jam in your car with the steering wheel in your hands at 8 pm on the way home, a shuttle from Google or Facebook flies along the fast lane, delivering employees home.
Thoughts? Comments?
