Comment by hintymad | Hacker News Reader

Comment by hintymad | Hacker News Reader

Parent Full thread

L7+ IC roles are not bad at all. Competitive packages. Tons of responsibility and freedom. I can't stress this enough. an L7+ really has lots of freedom and influence. They get to choose which meetings to go to, how much code they write, what architecture to use, who to work with, and have a serious say on what product features to launch, and which oncall to participate (except the GM escalation oncall). The company's policies and culture ensure that. They will be accountable for the architecture they choose, so of course they have the final say on what architecture to use -- typical freedom and responsibility. Plus, they have veto power of one's rating and promotion, after all. Other benefits include Lots of resources at their disposal. Good opportunity to learn from truly great engineers, at least in AWS. Note I'm not saying that every L7+ is great. All I'm saying that there are many truly great engineers and scientists that one can learn from. Think about the L7+ who built EC2, DDB, EBS, S3, SQS, and etc. Think about the L7+ who are fellows of ACM or NAE, who invented algorithms, built new systems, created new programming languages, and etc. They did not only spearhead the evolution of the underlying distributed systems, but also pushed large-scale application of queuing theories, formal verifications, and etc, as well as helped shape the engineering culture of the company. Oh, one also gets to learn the most elaborate and thorough operational practices. The production readiness review is amazing and is a gem for anyone to learn from.

Sure, being the top 1% of employees (which I'm assuming L7 principal is) at any company is sure to be great. Very few engineers will ever make that position at a FAANG.

Good news. L8 is now the new 7, thanks to rapid promotion in Amazon in the past few years, so the ratio is probably 3%, give or take. Joke aside, it's a fact of life that resources tend to concentrate to the top of a large company. For instance, partner engineers in Microsoft also enjoy great life. The real good news, though, is that wealthfront's CEO already gave actionable solution: join a blow-out small-to-medium company. The rationale is simple: what matter is growth. With growth comes challenging problems, career opportunities, talent density, and potential financial reward. That is, don't join FAANG, find a younger future FAANG. Of course, it's not easy, but it is definitely actionable and viable.

> For instance, partner engineers in Microsoft also enjoy great life.

Sort of. Partner engineers got hit hard with forced retirement packages in the first wave of interest rate hike fueled layoffs.

That's a different topic. A large number of E8/E9 in Google have been pushed out recently too. The key challenge is continuous growth. High-level ICs tend to become gatekeepers as the growth of their orgs stagnate. What they don't realize, though, is that gatekeeping accumulates debris and resentment over time, and the value of gatekeepers diminishes over time. A high-level IC either has to make things happen every few years, or reset their project by joining a new org, which means risking losing their institutional know-how.

I have not seen any L8/L9 ICs in Cloud get pushed out. If anything, G needs to hire more high level people.

I don't know about Cloud, but I saw lots of L8/L9s get laid off in CorpEng. Basically, anyone at those levels who didn't have a significant level of scope (i.e. enough FTEs under them), and even some who did, got laid off. I don't know exact figures, but you're probably looking at at least 50 FTEs for an L8, and over 100 for an L9, as a minimum bar to not be perceived as an unnecessary level of upper middle management that can be flattened away.

I was going to say, at Google I don't think the majority of the things the PP claimed as true at AWS apply until L8 (at least in Google Cloud).

Amazon has a flatter structure than Google, and it used to be that L7 was the highest rank (an internal doc cautioned the newly promoted L7s to stay modest even though, and I quote, "people may treat L7s as demigod").

>> blow-out small-to-medium company

This is surprisingly more difficult than it seems. You could go by VC fame, but even those have only a 10-20% hit rate, with a decent chance you end up at Juicely or whatever.

You could go by VC dollars raised, but that often sets you up for sales-driven companies rather than true engineering-rich cultures.

You could go for obvious stand-out products (OpenAI, Claused) but you notice there arent that many positions except in rare cases.

Am I over-simplifying this?

Yes, making sound choice is hard, just as our life often hinges on a handful of choices. That said, a coworker of mine mentioned an algorithm that seemed to work: first you pick a company that solves a problem that fascinates you and that has people are you willing to work with. You can check LinkedIn for the employe profiles, and you can rely on friend recommendations. The bottomline is, you minimize your chance of regret when it comes to choosing a company. If the company folds, at least you get to work with interesting people, build friendship, learn something new, and work on something fun and meaningful. The second second step is to rely on math. Give the company a time box, say 2 years. If the company does not have satisfying momentum in two years, then jump ship. Note it's about growth and momentum, however you define it. Everything else is secondary consideration. In 10 years, you will have 5 times. Now, the chance that you will succeed at least once 1 - P^6, where P is the failure rate of the companies. Even if the failure rate is 30%, you success rate will be 99.76%, an almost guaranteed success. And now, if you always join a top-tier company in the corresponding sector, your chance of success will even be higher. There are two catches, though. First, you need to be able to switch job whenever you want, so leetcode hard and study hard in your focused domain of experience. Second, 2 years of experience hardly can be enough for progressing your level of expertise, so you will have to work like hell to make 2 years of experience worth four years, instead of having one month of experience repeating 24 times.

Wealthfront's former CEO gave this advice. Ironically, even though he was a long term investor at one of the Valley's most storied VC firms, he himself did not run a breakout company. Instead, Wealthfront flamed out after 14 years with a mediocre acquisition at the peak of COVID startup mania.

Presumably, anyone at a staff+ level position in big tech, is likely influential and their word carries weight. So, I'm not sure how this is a pro or con for joining Amazon.

If anything, someone at that level earns so much money, and has so much unvested stock waiting for them, that even if they grow tired of the work, or disagree with its direction, they are unable to leave for something more fulfilling.

Being forced into the office 5 days a week might be enough to force you to reconsider though?! Maybe?!

the "unvested stock" thing is largely misunderstood. you have a target compensation number, and every year you get new awards to keep hitting that. if the stock goes up fast, your effective comp is higher than planned in the short term, but you get very few RSUs in the subsequent rounds. unless the stock goes up like a rocketship for many years in a row, your comp converges to slightly above target in the long term. although it has a strong psychological effect on some people, it's usually not rational to wait around for unvested shares.

That actually sounds kinda nice! (assuming you make it to L7)

> Think about the L7+ who built EC2, DDB, EBS, S3, SQS, and etc.

Does the average L7 person architect those services significantly, or just kinda maintain them? It's almost crazy to think about any old AWS employee (granted L7 is up there) conceiving those things, they've had such a massive impact on the ability to build things on the internet.

I'm not sure about those particular services, but in general the ground breaking and foundational things at the FANG and adjacent companies were convinced by people at much lower level than L7. It was a lot of L5-6 folks who then went on to be 7+ many years later.

Reply on news.ycombinator.com