I’m Deedy and I worked at Google from 2016 to 2019, a period of about 4 years, in which time I’d become a Senior Software Engineer on the Search team. My Google journey began in 2013, as a starry-eyed intern on the YouTube team taking pictures outside the dessert-shaped Android figurines in Mountain View (back when Jelly Bean was the latest version). Remember the Internship (2013), anybody?1 That released during our internship! Six years and many free lunches later, my time at Google took me all around the world — two and a half years in New York, at least six months on and off in Tel Aviv and most recently, a year long stint in Bangalore. In that time, I’d worked on everything from horizontal Machine Learning techniques for Query Understanding in to building out a large part of the Cricket product on Search, Assistant and the Now (Discover) feed. I even hosted an intern who joined Google fulltime a month ago!

At around August 2019, I was pumped to join Waymo, an adjacent Alphabet company that works on self-driving cars, on the Planning team as one of the first few employees in the Google New York office. It was a dream job. I was excited to be back in my favorite city in the world. But something didn’t feel right, and in a quick, unforeseeable series of events, I decided to leave to do something else.

Why I Left Google2 posts have become something of a tradition on HackerNews. I’ve read at least 7 of these, and I’d be remiss to not write my own. I write about the best parts of Google, the worst and what I’m going to be doing next.

The Good

Frankly, I don’t think I need to go over all the good parts of Google. Although they’re the butt of a lot of criticism today, Google’s impact on the world is undeniable. They’re one of the four largest public 3 companies in the world by market capitalization. Although now commonplace in the tech industry, Google was the company that championed putting their employees first. They championed engineering-lead organizations. They introduced the free food, the playful offices and the open culture many of their industry peers have emulated today. At its core, thye’ve largely achieved their mission - to organize the world’s information and make it universally accessible and useful.

Search had always been the flagship product I wanted to work on at Google. It is arguably the world’s most important internet product of all time. Such was my reverence for the product that one of the main reasons I applied to Cornell’s Computer Science program as an undergraduate was because of it’s strong information retrieveal program. Amit Singhal, former SVP of Search, had finished his PhD at Cornell in information retrieval under Gerard Salton and Claire Cardie, the latter who I was fortunate enough to take the Information Retrieval class with! Having the opportunity to sit in launch meetings with some of Search’s bigwigs was a dream come true - the likes of Pandu Nayak, Ben Gomes, Yossi Mattias and Shashi Thakur. They might not be your average celebrities, but many of them were mere folklore to me.

Google’s a great example of a company where “you either die a hero, or you live long enough to become a villain”. Amidst the current privacy debacle, talks of monopoly, and a slew of other critical press, a company that was a David 15 years ago is now a well established Goliath. What’s often lost in this dialogue is just how incredible the products Google have launched are. Do you remember that feeling when Drive and Docs made sharing and saving files obsolete? Or when you discovered that the internet could tell you how to get from A to B? Or when you didn’t have to use Internet Explorer any more? Google has become ubiquitous with it’s nine (!) 1 Billion+ user products4 and have left their mark on the world more than what a dollar value can measure. It’s impossible to imagine life without them.

And it goes without saying, Google has been a part of many of my cherished memories - long lasting friendships, incredibly talented peers, exceptionally good mentorship, and of course, the ability to work on products which touch almost a billion users.

The Bad

As a disclaimer, it’s difficult to generalize the negatives in a large, broadly successful organization, because the faults you see may just be happening around you. Broadly, from what I’ve heard, Search is regarded to be one of the better orgs at Google. Moreover, a lot of these criticisms aren’t exclusive to Googlen — they’re broadly applicable to all big companies. Nothing I mention here is really a secret, they’re well-known public facts in the tech community. That said, my biggest qualms were5:

  • Speed of Execution “If I wanted to work hard, I’d go to Uber” was a joke I heard a lot at Google. Engineering at Google is considered by many to be a retirement job, particularly amongst levels lesser than L66. Employees regularly rave about the work-life balance at Google compared to, say, Facebook, Uber or even Amazon. It won’t take you long at Google until you hear the phrases “rest and vest” or “coasters” - typically L5s who don’t want to work too hard for that L6 promotion (often regarded as the hardest jump at Google) and just want to get by working the bare minimum while raking in that meaty $350k7 wage. You’re not actively expected to get a promotion past this point. There are very few people around you are really pushing themselves to get stuff done, often prioritizing their own professional growth instead.
    Search is also a 20 year old mammoth codebase with its fair share of technical debt. Making small changes and evaluating its impact is cumbersome. Small one line code changes can have massive downstream consequences. Rigorous quality evaluation is, as it should be, essential to a product which is a gateway to knowledge for billions of users around the world. Unfortunately, the engineers suffer. Getting substantive projects done is often bottlenecked by technical debt, evaluation processes and dozens of often necessary red tape.

  • Professional Growth Until recently, L5 was the point after which you’re not actively expected to grow at Google. Now, it’s L48. Crossing that barrier to L6 is getting more and more difficult with time. Google’s promotion process, like any corporate promotion process, is flawed but frankly, I wouldn’t be able to suggest many ways to do it better at scale. Some argue that L5 should be split into two, like Uber does it. Some even think the granularity of levels should be much more fine, like Microsoft. At any rate, nowadays, getting to L6 takes at least 3 years in the best case and around 5 years at the median. Many factors are out of your control. In those 3-5yrs, you need to stick to one team, count on that team not being re-orged, hope your team’s function is a provable company priority, have a supportive manager and ensure your work is not only “L6-level” impact, but “L6-level” complexity. Often, you need to start managing a team by the time you get to L6 which means your promotion relies on headcount on your team increasing. Just doing what is best for the team consistently is often not enough to get you across the line. Many people have remained at L5 for over a decade, and not for lack of trying. Many say that getting to L6 requires significantly more work for a bump in pay that doesn’t materially change your lifestyle. More importantly, it’s not even clear whether you’re doing your best work either - often, you’re just gaming the L6 promo bar.

  • Unclear Monetary Upside At Google, like I said, you’re looking at 3-5yrs of extra hard work for a chance at a 40% increase on your compensation. At Waymo, there was even less clarity around compensation. Waymo structures it’s compensation like a startup within Alphabet. They grant their own options and RSUs of Waymo stock.9 These are illiquid until there’s an exit or a liquidation event. Given that Alphabet is the sole stakeholder, this means they’d buy back the equity. That is, you don’t get public Google stock, you get illiquid Waymo stock. When you move to Waymo laterally from Google, you maintain your level, your base salary and your existing equity continues to vest as it. Instead of additional Google grants, you get Waymo options and RSUs which is just unsellable paper money for now. For me, that would’ve meant about half of my total compensation would now be illiquid, but with more upside (and downside) potential. It just wasn’t clear how to estimate that upside in the best case. Was it gonna 10x? What was the value now? Could it 100x? Without knowing the current valuation, who knows.

  • Technical Growth Google isn’t the best place to learn transferrable skills. Until Tensorflow, Google really screwed up their strategy with infrastructure. They neither packaged it as a great cloud service nor did they open source it to the broader community. As a result, all these amazing internally developed pieces of infrastructure were pseudo-replicated externally10, subsequently gaining mass appeal from the wider technology community. That means all of the deep technical understanding you gain of infrastructure within Google doesn’t scale very well outside the company. Sometimes I ask myself - if I had to scale up a new product outside Google from the ground up, could I? And the honest answer was, I don’t know. My AWS knowledge was average at best. I hadn’t deployed a service on anything other than Borg (an internal Google tool) for almost 3 years. I wanted to grow my technical skills beyond just within Google.
    Secondly, slow technical growth at Google tightly relates to slow speed of execution. Your technical learning is proportional to how much you do, and if you just can’t do that much, how much are you going to learn?

  • Company Growth One of the joys at being at Google was the opportunity be a part of a selective set of highly talented people. Google’s grown to over twice its headcount during my tenure. Size has consquences. Talent dilutes, perks dwindle 11, and unity withers. Over time, no matter how much you fight it, you feel more and more like a cog in a gigantic wheel.

  • High Level Strategy Simply put, Google tends to play it safe, and people often prioritize personal growth over doing what’s right 12. I contend that this is because with size, equity ceases to align interests. As a company grows and diversifies their product offerings, the link between a person’s work and stock price breaks down. As a result, people don’t care about doing truly good work, but rather work that can be explained away as good with numbers 13. The distinction is fine, but important — a “Tragedy of the Commons” of sorts.

What’s Next?

I had quite a few options open to me when I was looking to leave Google, aside from Waymo — switch Google’s Product Management ladder, move laterally to Stripe, lead engineering for a startup in India, co-found a company with a friend, join a high frequency trading company, or to join the founding team of a company in the Bay Area. Each of these, of course, spanned a wide range of geographies, compensations, risk profiles and had its own repercussions on my immigration.

Eventually, I chose to join the small team at a Stealth startup in Palo Alto as one of the founding engineers. Unlike other startups I’d looked at in the past, the Founder had run a large unicorn in the past and was an elite engineer himself. The team was exceptionally strong, with veterans from the prominent companies in technology. I can’t publicly reveal what they work on just yet, but the product definitely has its fair share of technical challenges. It’s going to be an opportunity to learn from engineers much stronger than me, build out systems from the ground up, while still having a reward upside that could be life changing. I’m nervous, and I’m excited. I’ve never worked at a company so small, and I’m sure it’s going to be one hell of an experience.

Google is a difficult company to leave. Paraphrasing one of Paul Graham’s time-tested quotes, the more labels you attach to yourself, the dumber you become. I, of course, had fallen prey to this. Having spent most of my time at Google outside the Silicon Valley, even the so-called mainstream perks of Google were things I appreciated - the free food and snacks, the flexible work hours, the game rooms at work - all the bells and whistles. No matter where you’re traveling in the world or what kind of people you’re speaking to, Google is an immediately familiar name that catches people’s attention. That pride is difficult to get over. I still have a lot of love and respect for Google, and who knows? Maybe someday I’ll be back!

Footnotes

  1. Fun fact: the only other intern on the Youtube Captions team that I was on was James Damore. 

    1. Why I left Google. What happened to my book. What I work on at Facebook. Paul Adams July 12, 2011
    2. Why I Left Google James Whittaker, Mar 13, 2012
    3. Why I Left Google Spencer Tipping May 30, 2012
    4. Why I Left Google Ellen Huerta Aug 17, 2013
    5. Why I Left Google to join Grab Steve Yegge Jan 24, 2018
    6. Why I Quit Google to Work for Myself Michael Lynch Feb 28, 2018
    7. Why I left Google and Silicon Valley Christopher Johnson May 14, 2018

  2. Edit: Added “public”, duly pointed out by a friend

  3. Google Photos passes the one billion user mark, ninth product in Google’s roster to do so 

  4. Admittedly, there was at least one more fundamental reason for leaving that I’ve chosen to omit. 

  5. Details about levelling can be found at levels.fyi 

  6. These are public numbers obtained from levels.fyi 

  7. HN Comment 

  8. Blind - Waymo options vs Stock (WMU) 

  9. Github repository mapping Google tools to their external counterparts 

  10. Google will no longer hold weekly all-hands meetings amid growing workplace tensions 

  11. With Google’s distribution engine, any product you launch will get to 10m users even if it’s half-baked and doesn’t really work. After they launch, the entire org gets a promotion (“what growth! 0 to 10m!”). Next cycle, the product shuts down and there’s a max exodus of people to another team. Rinse and repeat. 

  12. Lies, damned lies, and statistics. 

I love hearing feedback! If you don't like something, let me know in the comments and feel free to reach out to me. If you did, you can share it with your followers in one click or follow me on Twitter!