Living Flat for a Decade, Yet Reaching the Chinese Ceiling: Wei Hui on Her Three Foundational Traits

The fastest way to learn something is to be thrown straight into that environment, where you realize people's expectations of you are right there, and you naturally start to act at this level. It's not like, "Oh, wait for me, give me half a year and I'll get there." No, you are thrust right into it. You have to be there whether you are ready or not, and you have to figure it out even if you don't know how.

In my own life, I've worked at big tech giants, startups, and traditional large corporations. Honestly, my least favorite are the big tech giants, even a company like Google. Even Google. Because it's a place that severely limits you. Within such a massive product framework, you only get to work on a tiny feature, and I just found that way too unfulfilling. If you try to do more, people around you will complain.

Deep down, I'm a pretty competitive person. I don't want anyone to consider me optional— someone who can easily be replaced. I just can't accept that. I don't have to be your boss, and you don't even have to view me as your best employee, but there must be something that nobody else can handle that I am able to solve. You have to become the go-to person.

On the West Coast, I feel like there's very little of that rigid hierarchy. Big bosses don't take themselves too seriously. The CEO of McDonald's, sure, he has his own private jet and close protection, but with security guards right by his side all the time, how are you even supposed to work like that, right?

I once went to India to build a GCC (Global Capability Center). My feeling about the situation there was just deeply heartbreaking. Once you step outside that corporate park, it's a completely different world. People are genuinely very poor; they don't even have houses, just plastic rain tarps where whole families live underneath. There are stray dogs everywhere, little puppies being kicked around or run over. Human life simply doesn't exist in any meaningful way under those conditions.

Hello everyone, welcome back to Class Representative Lizheng. Today we've invited senior alumna Wei Hui. Would you like to introduce yourself to everyone?

Hi everyone, I'm Wei Hui. My English name is Wei Manfredi. I came to the United States back in the 1990s to study. Early on, I came to pursue a PhD in Asian Studies. Uh-huh, humanities. Later I switched to Computer Science and stayed here.

You switched to CS while you were still in school? Yes. Because I graduated from Zhejiang University, and after graduation, I stayed on to teach. But staying on to teach, as everyone knows, if you want to get promoted, you must have a PhD. So at the time I thought, hey, the university is pretty nice, but I want to go out into the world and get a PhD before coming back. [Laughs] And I lasted half a year!

Wait, why? Why did it change? Actually, it's quite interesting. Because I was studying Asian Studies, right? My specific research focus was actually poetry. My advisor was from Taiwan and had stayed in the U.S. He asked me out of curiosity, "Do you genuinely want to pursue this academic path, or do you have other ideas?" And I said, "Even if I do genuinely want to pursue academics, so what?" He said, "If you really want to do this kind of scholarship, Asian Studies— this isn't the place to study it. Why don't you go to Europe? Northern Europe does it very well. Asian Studies, or Sinology, European countries were doing great work in that back then, places like the Netherlands, as well as Singapore, Taiwan, and Hong Kong." So he said, "If you truly want to do serious scholarship, go there. But if you just want to explore the world, right, and see what's out there, then I suggest to you, forget about studying this entirely— just switch to coding!" [Laughs] Great advice!

So after that, I thought, hey, that makes sense too. Then I went and asked my classmates. That was back in 1996. That's early—you guys were either just born or not even born yet! I was born in '88— oh wait... [Laughs] Okay, I am a bit older. But I'm sure at least half of the audience wasn't even born yet. Not born yet! Right, back then I was in the Midwest, and you just didn't see Asian people around. I was probably one of the rare few. So I asked everyone, "What should I do? If I don't study this, what's a good alternative?" A lot of people gave me advice: "You should just go into accounting. Because accounting offers very stable jobs, and you don't even have to talk to people." Because back then I was extremely introverted and didn't talk much. Oh really? You can't tell at all! Very, very introverted. Another suggestion was Computer Science: equally stable, and you don't have to talk much either. In my heart I thought, that sounds pretty good, so I chose Computer Science.

Putting that into the historical context, Netscape was around '97, right? Yes, yes. Right, that was right when PCs and Netscape were just taking off. Yeah, exactly. I just remember when I was in school back then, getting online required that specific sound: Beep... di-di-di-di-doo~~~~ That sound is still stuck in my head! Yes, absolutely.

So back then, I stayed and started my career. My very first job was actually at a massive engine company called Cummins Engine. You've probably never heard of it. In the Midwest, the entire town—the whole town—belongs to them, huge. They make engines for heavy-duty trucks. Wow, they make all of them. Okay. My job was embedded software engineering, basically doing simulations for them because they needed to test engine performance. Oh, so you wrote embedded programs to put inside the engines? It's somewhat similar to mechanical engineering. Yes. But the most incredible part was what? The most incredible part was that I had never even touched a car before in my life, right? I didn't even know how to fix a bicycle! [Laughs] But I became an expert. I could take apart an entire engine and put it back together within two hours, running simulations on that engine.

Right. And after that? After that, I felt I couldn't stay, because the internet boom had started by then. Just like you mentioned, Netscape and everything else was kicking off— the Web, web pages, right? It felt a bit like the AI boom we're seeing today. Exactly, it really was. So I made a pivot and jumped to a consulting company. Consulting firms specialized in working with cutting-edge technologies, diving right into them. So back then, I was working with ASP pages, which was brand new. At that time, you probably have no concept of what that was. None at all. Then a couple of years later I'd jump ship, another couple of years later I'd jump again, and eventually I made my way to the Bay Area in San Francisco. Back then, that was the real gold rush.

About when would that be? Around 2000. Hahaha! Right? Being in the right place at the right time. Exactly. Yeah, I think the environment is so important. Right. Geography is so important. Location, yes. That phase was probably extremely influential for me. Even though I didn't strike gold, right? Every company promised millions in stock options, but not a single one succeeded. I just didn't have that luck, I guess. Just didn't have the luck. But what did I learn from it? First, when you have an idea, you just have to go do it. And another thing that I think was even more valuable: I was never afraid again. A lot of kids, a lot of Chinese students coming over here, are so afraid of losing their jobs. Right, afraid of losing what you currently have. Yes. And because I kept job-hopping— especially after moving to the Bay Area— I changed jobs almost every six months. Because back then, if you didn't hop, people would come poaching anyway. Everybody was job-hopping. So after you hop, after you've switched a few times, you realize it's really no big deal. There's no such thing as "losing it all." You can find a new job anytime. You can find a new job anytime. And every time you move to a new place, you get a bit stronger. Move to a new place, get a bit stronger. I often tell my kids, there are a few things in life that are crucial, and one of the most important is being fearless, right? Not being afraid. You just think, what's the bottom line? Where's the rock bottom? And once you've hit that rock bottom a few times, you aren't afraid of anything anymore. That is so important. That's actually very true, right? I remember, when I first started working, I was so afraid of making mistakes. Yeah, and sometimes I'd ask a senior colleague in the team, "How can I do this without making mistakes?" Mhm. He said, "Go ahead and make a mistake for once." Exactly. Right? I didn't understand his advice at the time, but I got it immediately afterward. Right? Yeah. Right. So people think, "Oh no, I've lost my job, that's the bottom line!" But it's not the bottom line, is it? Mhm. What else can happen? Mhm. So back then, it wasn't just a mindset shift— becoming fearless and knowing your own worth, right? Mhm. In terms of working skills or technology, what kind of leap did you make back then? Every six months when you changed jobs, you were doing different skills, right? And when you change jobs, you surely want to step up a little bit, so usually you climb level by level. You force yourself to reach that level, right? How did you force it? By doing tasks beyond your projects at the company, or did you study on your own? Studying is a given, of course. I just fake it till you make it, right? Even if you don't understand, you pretend you do. But personally, I think the fastest way to learn is to throw yourself into that environment, and then you find out what people's expectations are of you right there, and unconsciously you act at this level, right at that standard. It's not like, "Oh, give me half a year, and I'll get there." There is no such luxury. So you are thrown right into it. You have to be there whether you're ready or not. You have to know even if you don't. Right, so you instantly find yourself right at that level. That makes a lot of sense. We're all human. Actually, in that kind of situation, you just didn't know people could work that way. Mhm. Once you know it, you adapt very quickly. You adapt very quickly. Yes, and it's also where you learn the fastest. When you're among peers at that same level, you quickly realize, oh, where is the gap? Mhm. And then you can quickly fill the gap, right? But a lot of people can't do that. Maybe they don't think others are that impressive. First, they can't even see where the gap is. Second, even if they see it, what they feel is frustration, or like, "Oh, I'm not good enough! Oh, I'm so scared! What am I doing in a place like this? Let me go back. Back to somewhere safer." Having this kind of mindset toward learning is already crucial in itself. Right, but the question is how you managed to keep up so fast. Mhm. I'd like to dig a little deeper: how did you spot that gap, and how did you close it so quickly? That's actually interesting. Let's talk about how to spot the gap first. How to spot it? For me, I am a very competitive person. I guess you are too, right? Most of us are pretty similar. What does being competitive mean? For example, there's someone in your team, say, who expresses themselves exceptionally well. Things you wanted to say but couldn't articulate, they just put it all out there. Wow. Then you think, "Hmm, that's interesting. Why were they able to say it like that? I was thinking pretty much the same thing, but I wasn't nearly as precise." Mhm. "They explained the whole thing clearly in just 5 minutes." Then you start wondering— how on earth did they organize their thoughts? Mhm. So that's one thing. After observing something like that a few times, you can repeat it yourself. Oh! Right? Most people don't seem to think that way. Most people wouldn't think, "If I watch a few times, I can replicate it." For example, I watch Obama give speeches—he's so fluent, and so inspiring, just speaking off the cuff. If I say, "I'll watch Obama a few more times and I can do that too"— no, that gap is way too big! Ah, that disparity is just too huge, right? But when you look at the people right around you, people in your everyday team, the difference isn't that astronomical. Someone who is slightly better than you isn't infinitely better. Understood. So you can keep up; it's achievable. Right, achievable, right? I used to deeply admire a colleague. I remember back in the early days, he sat right next to me, and he wrote code extremely fast. His approach was, "I'll just write it first, and then I'll test it and fix whatever comes up." Whereas for me, it was like, "Oh, I need to write it until I'm completely satisfied with it myself, and then I'll test it." Mhm. Back then, in the early days, we didn't have concepts like Test-Driven Development, right? Programming. So we would write things out pretty thoroughly, until we felt quite satisfied ourselves, and then we'd submit it for code review, and only test after the code review. But that colleague sitting next to me— he actually came from a music background, ah! He didn't follow that playbook. He hadn't studied formally; he wasn't formally trained. So he would write a block of code and test it immediately. Test it, then modify it. And I realized he was actually super fast, much faster than us. Seeing something like that—"Oh, he can code fast, so I can code fast too. Let me try his trick out." A lot of times, I think it boils down to this ego issue. People's pride gets in the way. "I actually do a pretty good job. I don't like seeing others do better than me." If that mindset is there, it will actually hinder you— just like you were saying earlier.

It's about having a strong competitive drive, but one that is paired with a lot of humility or what you might call intellectual honesty. Right? Being able to see other people's strengths. Yeah, combining those two things is pretty important.

Absolutely, because many people who have a strong competitive drive just turn it into rivalry, and that's the most unnecessary thing. Well, I think rivalry has its pros and cons. Let me give you an example. Yesterday, a student in our community built three projects at age 17, and all three projects failed. But he did a very honest retrospective, and actually, in every single project, you could see him growing really fast. However, I noticed a problem with him and brought it up to him—and he was super grateful for pointing it out. The issue was that he was using grand narratives to cover up some of his own tactical laziness. This is something competitive people easily fall victim to. Mm-hmm. He worked extremely hard, totally grinding away, but because he was so competitive, he wanted a narrative where he was better than everyone else. Oh! So instead of learning those hands-on skills or truly getting better at things he wasn't good at, he preferred to lean into and overplay what he was already good at.

Hmm, interesting. Because you've already learned so much and have such a rich background, in those areas, you certainly have a lot of strengths. Yes, but you still constantly work on your weaknesses to make up for them. Exactly, same here. This mindset is actually quite interesting. I've also met quite a few colleagues, especially Chinese colleagues— since we speak Chinese, we often eat lunch together— and I'd hear them say things like, "This person is no good, that person is no good." They'd make all these judgments, looking at people with a very critical eye, right? Well, everyone has flaws, but on the flip side, some people are very good at seeing others' strengths. They think, if I can learn from them, won't I become even stronger? Right? I'm the one who benefits! That kind of mindset is crucial.

Truly. And also, you want to choose friends who share a similar, more positive mindset. That way, when you notice a strength, oh, I noticed it too because you pointed it out, right? Yeah, I'm used to looking at other things through that lens as well, and I think that might be more helpful.

Mm-hmm. I think that's actually a talent— the idea that "if three people walk together, one must be my teacher." But being able to learn the right things with a similar mindset... often, you might subconsciously filter who you choose to learn from. How do you find the right mentors? That's possible too, because I've always thought of myself as a very humble person for two reasons: first, I'm very optimistic; second, I'm very positive. I'm humble, I easily see other people's strengths, and I've never felt that I am much better than others.

Part of that humility is probably due to personality, and part of it is because I made a career switch. Unlike you guys who— No, I studied economics, a PhD in economics. Ah, you studied economics! Many Chinese kids actually start programming in middle school. Well, I also programmed in middle school. Hahaha, right? I see. So at that point, they might have an identity— a sense of self-worth tied directly to that. Exactly. Their evaluation dimension collapses from multi-dimensional to a single dimension. Exactly, and at that point, you must be better than others. Yes, because a single dimension makes it easy to rank people, whereas multiple dimensions make comparison impossible. Multi-dimensional is impossible. Well said, that makes total sense. Multi-dimensional evaluation is so important. Mm, yeah.

Well, you've had a huge bubble over there— while it didn't instantly bring financial freedom, it gave you a lot of personal growth, helped you build a great mindset, and instilled something invaluable: being Fearless. Earlier we didn't mention what Huihui's exact title at Google was. What was it? Her title was Chief Architect for Retail Cloud —vertical cloud, Chief Architect. After that, she was VP of AI and Data at McDonald's, and now at IHG—the world's third largest hotel group— she serves as SVP of AI Architecture, Data, and Corporate Systems.

Wait, does IHG have a CTO? She's pretty much equivalent to a CTO. IHG doesn't have a CTO, yeah. Career-wise, this is basically one of the ultimate ceilings for ethnic Chinese in the US, alongside entrepreneurs like Eric Yuan (founder of Zoom) or certain Silicon Valley executives. But she has basically reached the top tier in traditional corporations. Yes, if you're on the technical track, reaching this level is about as high as it gets. Yes. So it is a very, very successful, outlier kind of career path. But as you said, perhaps in the beginning, it wasn't something planned out step-by-step. Never! I never planned to move into management. When I was young, early in my career, I strictly resisted going into management. I thought, first, you have no real skills; second, you just talk empty words; third, beyond a certain point, you become completely useless. That was my thought back then. What do I think now? It's completely different. Completely different. Yeah, I was just being naive back then, haha! Why don't you talk about it directly now? How do you feel about it now— what is the true value of management?

The value of management... is really what we always read about in martial arts novels when we were kids: "four ounces can move a thousand pounds." Leverage. It really, truly is. The absolute essence of management is leveraging force. You are mobilizing such a massive team, getting everyone's hearts aligned to follow you and do what you want to achieve. Your vision is out there, you get them to understand your vision, and then they execute. Thousands of people working hard, moving in the exact same direction— that is shifting a thousand pounds with four ounces. I think that is incredibly difficult.

Ah, I get it. Tech students tend to think that building the actual product is the ultimate achievement. But what you build is actually more important than how you build it. Because if you choose to build the wrong thing, making it is completely useless— it's just a waste of time, right? That's why vision matters. Second, how do you get everyone to understand your vision at their respective micro-levels? Because everyone takes on a small piece, right? And you have to piece the big picture together. So first, you make them understand the small piece they are working on.

...is one piece of this huge puzzle, ensuring it's executed with extreme precision. Second, all those disparate pieces need to be brought together—assembled with pinpoint accuracy to drive things forward. Yeah, that's definitely not something you can pull off just by talking a big game. In the abstract, people might think, "Oh, that's easy." But when it actually comes to execution, skill levels vary wildly. The problem is that these qualities are invisible— there's no metric saying, "This leadership scores this many points or that many points." Exactly. Right. So if you're solely focused on the purely technical dimension, you might not even comprehend it. Yet it's very real and crucial.

It really is. You only truly appreciate it after you've experienced it yourself: "Now that's a master." It's just like programming— technically, anyone can write code; anyone can finish a program, right? Anyone can complete a project. But once you look under the hood, you immediately see the quality. Right? Yeah. After it runs for a while, you realize their caliber— the architectural standard. When you add new features, you don't have to rebuild everything. Exactly, you can build it up layer by layer. That's where the design quality shines. So you only recognize what's high level versus what's low level once you've done it yourself.

That's why the number of people who can actually judge the quality of leadership is inherently very limited, because they themselves need to have that kind of background first. Well, some people just have natural talent— they can spot it at a glance. That's different. Someone like Liu Bang could tell immediately. (laughs) Haha. I remember an example, I think it was Shi Jingtang [founding emperor of Later Jin], an emperor surnamed Shi during the Five Dynasties and Ten Kingdoms period. He was reading the Records of the Grand Historian, looking at the different advice strategists gave to Liu Bang. When he read a second-rate suggestion, he said, "Oh, if he took that advice, he'd be finished." Then when he read a great piece of advice, he said, "He must follow this one." Oh! It matched Liu Bang's exact judgment. Oh, really? Yes, that's pure innate talent— right, a talent for judgment.

Please go on. So after Russell, I worked at a gaming company for a bit, then at Nordstrom for a while, and Lululemon was where I stayed the longest. Ah, I see. Oh, right, right, right. I remember seeing Lululemon on your LinkedIn; before that, I probably didn't know as much. Right. Before Lululemon, I also spent some time at Visa. At Visa, I felt like I learned an important lesson. A lot of people join big tech giants— and Visa is relatively speaking a very tech-driven company— and they think, "This is amazing," right? But it's vanity. Why? First, once you're there, you realize you're working on such a tiny piece. In a small company or elsewhere—bam— you have a huge scope; you can touch everything. Exactly, you can do the entire full stack. But there, you're confined to this tiny slice. I felt so suffocated— extremely suffocated. Within such a massive product landscape, you're only allowed to touch one tiny feature. Ah, it felt so unfulfilling. And you couldn't do extra, or people around you would complain. They'd complain? Yeah. And then you'd look underneath at the codebase, and it was completely terrible. (laughs) Because so much of it was acquired! Yes, exactly. They'd acquire small companies, right? And it was messy— just hack on top of hack on top of hack, piled up like that.

I really never understood it. In my own career, I've worked at big tech, startups, and traditional big corporations. Big tech companies are actually what I disliked the most— large tech firms. Really? Even a place like Google? Even Google. Because it heavily limits you. What you work on is so small. Many people might not realize how small their scope is. Sure, some get lucky, but for the vast majority in big tech, what they build is remarkably small.

I think people have an illusion: they assume that because what they build is used by millions of people, they have massive impact. But the reality is that despite all those users, you have zero decision-making power. Exactly, you don't have real impact. Exactly. With or without you, it would be built the exact same way. Yeah, you just take a task and... You're just fastening a bolt. Exactly! You fastened a single bolt on a 10,000-ton cargo ship, but at the end of the day, you're still just turning a wrench. Exactly, just turning a wrench. Right, that exact feeling.

Now I tell my kids, big tech experience is important because you need that confidence— you've seen the big stage. Right, you've seen the world and proved your capability there. But ultimately, you still have to follow your heart. People are just wired differently. I think there are two main distinctions I realized after branching out on my own:

1. Do you prefer certainty or uncertainty?
2. Do you prefer executing tasks or delivering outcomes?

Because when delivering outcomes, you have to confront uncertainty, hunt down all sorts of resources, and take full ownership yourself. Many people hate taking personal responsibility. They just want someone to hand them a problem, solve it, get paid, and receive a reward— do well and get an even bigger bonus. That's interesting. It really is. That's why people love big tech environments so much: "Hey, I just need to know with certainty what this task is, and know for sure that I delivered it." But I think that kind of work will disappear in the future. It will definitely disappear. It's already disappearing. AI is exceptionally good at task execution.

So after Nordstrom, Visa, and Lululemon... When did you start moving into management at Lululemon, or climbing up the ladder? It basically started at Nordstrom. At Nordstrom, I was a Principal Architect, managing several solution architects underneath me. How did you jump straight to Principal? Well, remember how long I had already been working by then! When I joined Nordstrom, I had already been working for over a decade. So progressing from engineer to senior engineer, and then shifting toward architects, right— Solution Architect, Senior Architect— it went through stages.

And there was another detail I overlooked earlier: when you said you basically "coasted" for over ten years, that wasn't really the case. You mentioned the work was exhausting; you just didn't have to travel. Exactly. Right, so the workload and intensity were actually quite high.

Under high work intensity, your technical skills are actually moving forward all the time. — Yes, exactly, right. I have a habit, and sometimes I tell my kids, "I don't know why you guys are a bit like me." What is my habit? Well, deep down I'm quite a competitive person. Although my life state and life philosophy is be water, just like water, you know? — Mm. But, there's one thing: I refuse to let others think of me as optional. Dispensable. I cannot accept that. I've never been willing to accept that. Even in a workplace, I might not be your manager, and you don't have to treat me like your best employee. I don't chase after that. But there must be one thing: everyone has to come to me. When others can't solve it, it has to be something I can handle. So every time I go to a new company, I look for this exact point. I have to become, you know, the go-to person. — Mm, this is very important. — Yes, this is extremely important. That way, you naturally get pushed to the next position. Pushed to the next position. I mean, from a technical perspective. Can you share a few stories? Let's share some stories, huh? Some things are very simple. When I was young, for example, using databases a lot, right? Fetching data, especially in an investment company, when you write queries, actually, there are many ways to write them so they're very efficient and run extremely fast, right? — Right. Even when you talk about joins— left join, right join, right? Everyone knows them. But joining has its tricks. And knowing which side to put an index on and which side not to also has its tricks, right? — Right. So if you read some books and practice this stuff, you can actually figure it out. So in the end, the query statements I wrote were the most efficient, the fastest, and everyone thought, "Oh, if my query isn't working and it runs too slowly, I'll just go find Wei." — Mm. So this led to everyone coming to me. Although this was a very small thing, what was the benefit? The benefit was that— my name, even though I didn't go around promoting myself, everyone knew this person existed. The scenario might be a group of people having dinner, and someone says, "Hey, my query is running so slowly." And someone else says, "You should go find Wei." — Exactly, exactly. "Just go find her." Right. Or say a problem occurs in the company, for instance, with a training flow, once the system is running and not returning back— which happens often, right? Long-running queries, and you can't just kill them. So what do you do? Call Wei in the middle of the night. This dramatically increases your visibility. Because usually, upper management is the one troubleshooting. Right, this is the most anxiety-inducing problem, right? So you become the firefighter. This is very important. You become an indispensable cog in the machine. Even though you're still a cog, you are an essential one. I think this is quite important. Even if you aren't actively climbing that ladder, you still have to become (that irreplaceable person). You have to nail your positioning, exactly. In fact, a lot of business competition is just like this. I've read many business books, like Blue Ocean Strategy and things like that. They say you shouldn't try to be the best; striving to be the best always leaves you with no profit. You must strive to be unique, and be indispensable in that space. That's right, being indispensable is crucial. The more indispensable you are, the higher your bargaining power. Yes, very important. I feel everyone really needs to pay more attention to this. You can actually see this if you think about it during your daily work. Even in my current workplace— I've only been at IHG for two months— but I've already noticed a few people and will specifically call on them. For example, there's someone who— it might sound funny, but they make really great PPTs (presentations). That matters a lot. Extremely important. They can take complex information and make it very clear, expressing it simply. When I was at Tencent, this was one of my core values. Right? People might think making good PPTs is just about aesthetics, but it's not. The primary purpose is intent: clearly conveying what needs to be done, and being able to drive critical decisions in key meetings. Exactly, and the interface of how all their information is presented is top-notch. Mm, so the boss listens and goes, "Okay, you've made this point clear." Then I can make decisions with peace of mind. This is actually a very important skill. You might ask, is this a hard-coded, hardcore technical skill? From a programming perspective, no. This person doesn't write code, but they do this extremely well. Right? When you talk to them, they immediately grasp your main points. You don't even have to explain, mm-hmm. And whatever you fail to explain clearly, they help you articulate. That is awesome, right? Just like a super-powerful large language model! Truly, yeah, mm. And another example: I just promoted someone to Director of Operations. They are exceptionally good at handling miscellaneous tasks— whatever you hand them, you can rest assured, and they can think one step ahead of you. Usually, we just follow up, right? "Here's the, you know, action item, this is what we need to do," and assign it out. But they anticipate you. They come and tell you, "These are the things you'll probably need other people to do, and these are the exact people who should handle them." They've already laid it out for me. That's pretty impressive. So I felt I should promote them. So, although these seem like small things, they make someone stand out significantly. I don't think they're small at all; these are all critical. But everyone's strengths are different, that's what I want to say, right? Some people are great at coding, some excel at handling miscellaneous tasks, and some are fantastic at making PPTs. Right? Mm. What is presented on the surface might be just one thing, but behind it, none of it is simple. Especially when you're already in such a high position, everything that lands on your desk is important. Right. So when you find someone truly useful, it's usually because the task itself is vital, requiring a very strong combination of skills. It's desperately needed, yes. Come on, keep going, keep going! What else? We're about done, right? "Now your career is just beginning, as a principal architect managing a bunch of solution engineers..." People are going to be bored to death listening to this! It doesn't matter, people actually enjoy hearing longer-form content. Wei, conclusions and summaries can be done very well by AI; rather, it's the stories in between— where people share their own judgments and thoughts—that are valuable. All right, then keep talking! Keep going, keep going. Then you got to Google.

If you look at your career trajectory from a regular developer all the way to where you are now as an SVP, plotted against time, what do you consider the most critical milestones?

The most important milestone was definitely Lululemon. When they hired me, it was for the role of Director of Architecture. At many companies, architecture is just architecture—not strictly a game-changer— but for Lululemon, it was crucial. Timing and luck played a huge part because the company was right on the verge of a major digital transformation.

Back then, Lululemon didn't have a website or any real online business, and very little international presence. It was a brand with a strong personality that relied heavily on community-driven word-of-mouth. They didn't have formal sales teams, marketing, or even ad campaigns. But their products were fantastic. When the board decided it was time to scale up, they parted ways with the original founder and brought in a professional CEO. A professional CEO comes in to drive revenue, so right away the mandate became: go digital and expand globally.

Once I joined, it became clear how critical architecture was because there was practically no foundation. In those days, their internal IT department mostly handled infrastructure, networking, or basic desktop support—like fixing someone's broken computer. There were no in-house software engineers; any small applications they needed were completely outsourced. So when you want to build an entire e-commerce business from scratch, you have to architect it yourself.

I ended up staying at Lululemon for nearly five years. Every new capability we launched seemed to double our profit margins. It was an exhilarating experience being right at the center of that engine. Because architecture was foundational, my scope expanded from architecture to strategic planning, aligning where to invest in tech with the architectural roadmap. When we realized the baseline data and AI capabilities were lacking, I took over the engineering teams to build out those functions.

Over time, my title evolved: first Director of Architecture, then Director of Architecture and Strategic Planning, and eventually Senior Director of Engineering and Architecture. They kept expanding my scope. As a Senior Director, I oversaw data and AI—which at the time was traditional machine learning focused on marketing, customer segmentation, and recommendation engines—alongside core architecture.

By that point, Lululemon had become a massive success story. Other companies wanted to study our playbook and poach our talent.

It looked magical from the outside: how could a brand grow that fast internationally, essentially doubling year over year? We used to joke about it: when I joined, the stock was around $40 a share; when I left, it was over $400. There’s that saying: "Even pigs can fly if caught in a whirlwind." Riding that hyper-growth wave makes you incredibly agile, gives you deep experience, and builds your industry reputation. For a while, whenever anyone needed a high-caliber architect, they would reach out to me—either to recruit me directly or to ask for referrals.

Even companies like Google reached out to poach me. Initially, I didn't want to leave Lululemon because I loved the company, but the opportunity at Google was too compelling. At the time, Google wanted to build out vertical cloud solutions, specifically Retail Cloud, with a suite of products like Retail Search, Retail Recommendation, and Retail Inventory management. They hadn't built them yet, but these were all systems we had already successfully deployed at Lululemon. Before Google, Procter & Gamble (P&G) had also made a strong push to recruit me as VP of Architecture and Innovation. I got all the way to the final offer, but thinking about the freezing winter location, I decided I couldn't relocate.

Google was appealing for two reasons: first, I didn't have to relocate—the role was based right in Kirkland; second, the product domain was something I was genuinely passionate about and highly skilled in. So I moved to Google and spent over two years there, right around the dawn of the current AI wave.

My role was essentially advising enterprise C-level executives on Google’s core technological strengths and how to apply them. While I wasn't in direct sales, it was aligned with the strategic side of sales— a consultative soft sell. I never pitched specific SKUs directly; instead, I showed them which architecture and solutions would solve their specific business problems.

During my time at Google, I was tasked with helping define "big bets." Big tech companies often shift focus quickly— if an initiative doesn't gain traction immediately, leadership pivots. We realized that standalone vertical retail products were struggling to gain broad enterprise traction and were extremely heavy to build. Google didn't fully appreciate the sheer complexity of enterprise environments: they had incredible core technology, but lacked understanding of the last-mile delivery required to take deep tech and make it fully operational inside enterprise workflows.

So we decided that instead of spreading resources across ten niche products, we would consolidate into four major big bets and try to scale those. I led four agile teams to develop those products, one of which was Google Distributed Cloud. That was the journey.

Right, exactly its application in retail, right? With distributed cloud, we didn't initially think about retail, but retail turned out to be extremely important. Why? Retail businesses have physical stores, and physical stores will inevitably lose their internet connection from time to time, right? When the network goes down, you still have to run your business. Whether you're running a restaurant or an apparel store, you still need your business operating and your POS system running. So distributed cloud actually became a crucial piece of infrastructure for retail.

Makes sense, that's really interesting.

On top of that, for most enterprises with brick-and-mortar stores, deployments were extremely complex—you had to go store by store to deploy. So Google Distributed Cloud provided a single control panel: you could deploy everywhere and have full visibility. It was all quite fascinating.

When we started working on it, McDonald's came along. Because McDonald's really wanted this Google Distributed Cloud, they tried hard to poach me. They told me to come over, so I transitioned to McDonald's for one main reason: when you build these things at Google, you don't really get to see your end users. But once you're at McDonald's, you see them directly. That was something deeply satisfying for me— that direct feedback you can immediately see. That was really important.

At McDonald's—I remember back then, we happened to do a show together, and you had to hire a lot of people at McDonald's. What was that experience like? Was there any culture clash?

Yes, it was pretty painful. I hope there's no one from McDonald's in the audience! I felt the cultural clash deeply. For instance, between a data-driven tech company and a traditional enterprise, the difference is night and day. In a data-driven culture, you build something and test it out in the wild first. Traditional enterprises are terrified of that. They insist that everything must be tested to absolute perfection before launching. It's a completely different mindset. That was the biggest clash.

The second clash was that decision-making was painfully slow because everyone was afraid of making mistakes. Lululemon is also an apparel company, not really a tech firm, but its entire culture was: "Hey, give it a try. Nobody's going to blame you if it doesn't work out, and you won't get punished." There was none of that—if something didn't work, you just quickly pivoted and swapped it out. That was totally normalized. But in a traditional enterprise, the moment you make a mistake, you take the blame and become the scapegoat. So that was a massive friction point.

Another clash was the regional culture difference with the West Coast. On the West Coast, people rarely dwell on hierarchy. So what if you're a big executive, right? The key is that big bosses don't take themselves too seriously. Everyone comes from similar backgrounds—either as software engineers or product managers. But traditional Midwestern companies are completely different. The boss is genuinely the boss— there is a strict class structure.

For example, our CEO at McDonald's had his own private jet, and if you needed to approach him, he was surrounded by security guards. How are you supposed to work collaboratively like that, right? When that cascades down level by level, the whole corporate culture becomes rigidly hierarchical. I found that environment extremely stifling for innovation and collaboration. The engineering culture just vanished, replaced by a management culture where managers make all the decisions and engineers have very little voice.

What felt most suffocating to me was that they were terrified of open debate. You couldn't debate anything—everything was either agree or shut up. That was it; no room for debate. Coming from an engineering background, we rely heavily on debate. True insights only emerge through debate; otherwise, how do you know what's right or wrong? So that was something exceptionally hard for me to get used to.

I had no idea McDonald's was like that. Their CEO recently had that whole burger incident as well. Once you understand that culture, it's easy to picture. It's actually quite common, especially among traditional, large Midwestern corporations. First, the top CEO genuinely wields immense power. Second, they have a deeply ingrained sense of hierarchy because they usually didn't start out as programmers or technical professionals who disregard hierarchy. They climbed the corporate ladder step by step, so their sense of status is very strong— quite similar to state-owned enterprises.

And after McDonald's, you moved on to IHG?

Yes, going to IHG was a similar story. I was already feeling somewhat stifled at McDonald's, and it was also genuinely exhausting. Why? I had to hire so many people—my team back then had over 1,000 people. To scale that, I needed to recruit a ton of talent, but trying to hire strong engineers was extraordinarily difficult. They wouldn't offer competitive, tech-grade salaries, and without matching tech compensation, you simply can't attract talent.

On top of that, contractors were exorbitant. We used Accenture contractors back then— easily $250 an hour at a minimum, and $350 for the better ones. It was insanely expensive. Yet when hiring full-time employees, you couldn't raise their base pay because of rigid salary bands. It was a real headache.

Later on, we thought: why not build a center in India? Maybe we could hire great Indian engineers without the costs skyrocketing. So we set up a GCC (Global Capability Center) in India. That turned out to be exhausting—you constantly had to fly out there. Plus, without an established leader on the ground, it was hard to spin the team up quickly. It was thankless, grueling work that ultimately wasn't worth the effort.

Furthermore, seeing the situation in India felt heartbreaking. That colonial dynamic is still visibly present. When major corporations set up operations there— much like what we saw in Shanghai— it feels like a colonial enclave: immaculate, ultra-modern, practically identical to the United States. Like that American TV show Outsourced from over a decade ago, where people worked in gorgeous offices, but the moment you step outside the campus gates, it's an entirely different world.

Yeah, there was trash everywhere, and people were living in deep poverty. Many didn't even have a real house—just a makeshift plastic tarp with an entire family living underneath. That stark wealth gap was unbearable for me. What broke my heart the most was seeing the streets completely covered in garbage and overrun with stray dogs— puppies being kicked around or run over by cars. In that environment, life was treated as if it held no value. It was deeply distressing. Going there every month and facing that reality repeatedly was tough. On top of that, as a woman, even the Indian government advises against going out on your own. So I felt severely restricted, confined to a tiny bubble. It wasn't the best experience, so I ended up moving to IHG.

What was your vision when joining IHG? What are you looking to build next? We have many viewers in tech— could you share what opportunities you see in hospitality, and in this era of AI, what kind of talent the industry needs?

I'd love to share why I jumped from McDonald's to IHG. First, hospitality is a fundamental necessity— people will always need places to stay. Second, if you look at the hotel industry, has there been any major innovation? Almost none. You can do self-check-in at airports today, but many hotels still don't offer it. It's tedious, right? If you need a bottle of water in your room, you still have to dial the front desk on a landline. Who makes phone calls anymore? Exactly, a traditional landline. It's such a friction point. Plus, even if you stay at a hotel ten times, they often have no clue who you are. Even at luxury hotels, you're treated like a stranger every visit. Yet when you go to any other high-end luxury business, they recognize you instantly because customer lifetime value matters. But hotels historically lacked that mindset. That made me realize this industry is primed for disruption— especially with AI, which can help it leapfrog ahead after lagging behind for so long.

When I asked people what innovations the industry has had, they said, "We have innovations!" Like what? "Well, first, platforms like Expedia and Booking.com came along and moved booking online." I said, "Sure, but everyone moved online." "What else?" "Oh, Airbnb!" But that's an external competitor—what does that have to do with hotel innovation? Exactly! [Laughs]

So at the end of the day, the industry is quite legacy and ripe for disruption, yet it enjoys very high profit margins. Look at their business model: the Big Three management groups—Marriott, Hilton, and IHG. What is their strategy? If you want to open an IHG or Marriott hotel, you fund it yourself. You invest $5 million or $10 million to build the hotel. They provide the specifications—wall standards, design guidelines—and you build it. That's step one: brand standards where they invest no capital, and you pay them franchise fees. Second, you use their software— often quite dated software— and pay them recurring software licensing fees. Third, they introduce you to a third-party management company. Many property owners have capital but don't know how to run operations, so the brand connects them to managers and takes another cut. The brands barely put up any capital of their own; it's an asset-light, brand-driven model. That's why their profit margins are so high with minimal risk. And that's precisely why the industry lacked the drive to innovate— they make plenty of money whether they change or not.

From my perspective, our leadership— especially at IHG—is very forward-looking. They recognized this as a rare window of opportunity: the industry hasn't made major strides in years, and as the third-largest player, IHG can't simply match Marriott or Hilton's raw property count overnight. We have over 6,000 properties, while they have double or triple that. The only way to leapfrog is through intelligent hospitality:

1. Smart, automated hotels: end-to-end automation drives operational and maintenance costs to a minimum. That makes franchising far more attractive because owner profits increase.
2. Personalization: learning from retail and entertainment to personalize the guest journey. For example, if a guest arrives with a child who wants to play while the parent needs to record a podcast, the system anticipates those needs: recommending nearby kids' activities, providing a babysitter, and preparing a quiet room to record. That level of personalized service is decisive. If you deliver that, guests won't mind paying an extra $40 or $50 a night over a competitor.

That is our core mission: making hotels fully intelligent and scaling hyper-personalization. Since joining, I've focused on two main pillars. The first is building digital twins: a hotel digital twin, a customer digital twin, and an operational process digital twin. On top of those three twins, we deploy agentic AI. Digital twins are already well established in manufacturing—companies like Ford have long used them to monitor real-time factory floor status. But historically, you could only observe, not act. With agentic systems, you can take automated action. If real-time monitoring detects a long line forming in a lobby, the agent can automatically dispatch staff to open an additional front desk counter. That is the agentic component— embedding active intelligence across the entire operation. If a guest was scheduled to check in at 9:00 PM but their shared flight data shows a delay, the system can instantly message them:

"We already know your flight is delayed. We've got your room ready. There's no need to rush to check in." That creates such a warm feeling. Even if you haven't experienced it directly yet, yeah, that's what I call anticipatory service. It's so important. Anticipating what services you might need, and then proactively preparing them for you beforehand. So getting that right is crucial. But what's the foundation? Having real-time signals is essential. Digitization, first and foremost, is critical. Digitization, right. Once you get that foundation set up, exactly, then you can build on top of it. So working on this is actually super thrilling. Why? Just think about it, we come from a data background, right? Over 6,000 hotels, all over the world. You build a digital twin for them, feed in real-time data, and process it. Second, you detect anomalies: like what should I pay attention to? Your book is probably about this sort of thing, right? Some of it, some of it, yes. And then diagnosis: you know, given this situation, right, the pattern — what was the past pattern and how did we solve it? You learn using this method, and then give your diagnosis on why this is happening. Followed by recommendations, and then actions. So these four steps you can do with AI on your digital twin. So building this is actually extremely concrete, right? It's hard and complex, yes, but also very concrete. Very concrete indeed. And then the results? You can see them immediately. This naturally becomes something we really want to build.

So does IHG give you the corresponding budget or the necessary authority? Are you still like back at McDonald's, only able to hire outsourced contractors, or can you hire some top-tier talent to build this? Let me tell you, this is fascinating. When they recruited me, AI agents and AI coders weren't this powerful yet. So the original plan was, oh, we'd hire a big team. But over the past three months, things have gotten so insane, as everyone knows. Especially in the last two weeks, a lot of companies in Silicon Valley were saying, "We now have 80% of our code being written by AI." If that's the case, I don't need a massive team right at the start. With that in mind, I'll just hire one senior engineer, maybe a junior one later, and let a bunch of agents write the code for you. So this turns into a squad. The old minimum squad size was the "two-pizza team," right? Now, you're just given two people, and you create everything else yourself. First, let's spin up a few squads and see how it goes. So this has actually become a very challenging endeavor. And I've always said, I'm not someone who needs a huge team or a massive budget. I actually thrive on constraints. I feel like if you give me constraints, and under those restrictions, I can still achieve what could previously only be done without restrictions, that's when I feel most proud. Exactly. You need someone like an architect or a conductor. Exactly, someone who deeply understands the business, moves very fast, and actually ships these things. So for skill sets, as you asked about who I'm hiring now, I don't necessarily look for people who have a ton of AI experience. I don't think that's very important. As long as you're a good software engineer, that's enough. You know what good looks like, you have that core training, and I think you can pick up AI very quickly. Especially now, because AI tooling is genuinely powerful, I really believe that. Especially when you want to build an AI product

that is extremely reliable. But on top of that, we've also found that there are some things you need to unlearn. That's right. For example, in our community, aren't we trying to translate all Chinese text into English? The traditional software engineer's approach is: Chinese API English. Then, what do you do when the AI starts getting lazy or hallucinating? You slap a bunch of engineering work onto the API to solve it— like text chunking, consensus mechanisms, and all that stuff. And as a result, it ends up being humans cleaning up after the AI, with more and more edge cases popping up. Exactly, exactly. Later, Brother Ya's approach was different. He realized, "This isn't an AI-native approach. The AI does some work, and humans constantly have to clean up its mess. Instead, it should be the AI cleaning up after the AI." How do you do that? For instance, don't use a rigid workflow; use files instead. Because file states are real-time, and the working state can be persisted. Then you describe to the AI what kind of file you want to transform this file into. You spell it out: "Okay, translate this into English, and here's what kind of checks you need to run." Once you clearly describe all of this, it figures out on its own how to get that file into great shape. And once the job is finally done, you just publish it. Ah, that's interesting. Right, it's a paradigm shift. Once you fully grasp that concept, it shifts from process-deterministic to outcome-deterministic. Exactly, focusing on the outcome. Right. And you realize that everyone's initial reflex is process-driven. Yeah, as a programmer, your first reflex is always process-driven. Right, but shifting to outcome-driven is something you have to unlearn. Ah, that is super interesting, yes. There are a few things like that. You find that traditional software engineering principles still apply, but at the same time, certain habits need to be unlearned. Exactly. So what's your advice? I'm currently building a team, what kind of skill sets do you think I should look for? I think first I look at mindset, and then skill set. Mindset and habits: are you willing to learn? If someone's self-worth is tied strictly to the quality of the code they write, then they will find it very hard to accept that AI can write better code than they can. Exactly. They need to be willing to learn, and they need to recognize, "Okay, this is something new, and I need to learn and master this new tool. I need to become AI-native to be a more formidable person." Whether they can accept that is the number one criteria, I think. Then, proactive learning, having intellectual honesty, and being humble. Right, yeah. When we are challenging past unchallenged best practices or underlying assumptions, they need to be able to look at the problem from first principles. Not just thinking, "Oh, this is the holy grail of software development, therefore it must be right." Instead, they should be able to say, "Ah, under the old environment, why was this right?" For example, a key observation Brother Ya made is that a lot of past software engineering best practices existed because writing code was too expensive. Right, so your code had to be scalable, and it had to be reusable.

So people always preach about modularity and best practices, but once writing code becomes cheap, a lot of that is no longer necessary. Exactly. But someone deeply entrenched will argue, "The reason I'm so valuable is because I've mastered these scalable practices." So challenging that methodology feels like an attack on their personal worth. That's why grasping the core essence of a problem, reasoning from first principles, and figuring out how things should be built in this new paradigm is so critical. On top of that, agency is essential— the drive to proactively learn and relentlessly push things forward.

Silicon Valley companies often say they need two archetypes of engineers: an architect and a pirate. A pirate does whatever it takes to hack things together and get it out the door. It doesn't matter if the code is a mess, as long as it achieves the business goal. Then the architect steps in to clean up the mess, make it scalable, and break it down properly. I see. Yeah, exactly.

Right now, my biggest challenge is that most of our existing team has been with the company for years. In traditional enterprises, people stay for a decade— sometimes 10 or 20 years. The hardest part is unlearning. Unlearning is the ultimate challenge. Yet it doesn't have to be; you have to create a sense of "magic" and use that magic moment to help them willingly embrace this massive shift. The silver lining is that the threat of AI feels very real now; everyone recognizes: "If I don't learn AI, I will be made obsolete." Exactly. Back when we were training teams in 2024, many would push back: "My code is better than AI's; I don't trust AI." Teaching it this year, nobody says that anymore. Right? People don't think like that anymore. But we still explicitly point out why they might feel resistance. Once you surface that in a collective environment where everyone shares the same realization, people's mindsets genuinely shift.

Once the mindset changes, the rest falls into place. The technical skills are actually easy to teach. Once engineers shift gears and begin using AI to produce 99% of their code, they inevitably run into roadblocks. Without a shifted mindset, they'd blame the tool: "AI is unreliable; AI can't do this." But once their mindset has transformed, they think: "The issue is that I don't know how to prompt or guide AI properly. Top performers are already building incredible things with AI; if I can't, it's because I haven't mastered it yet. What am I doing wrong?" That opens their mind to learn. And through learning, they realize: "Oh, I see— my evaluations weren't defined properly, I wasn't setting up my pipelines right, or my underlying architecture was flawed."

That also connects to your recent article, which resonated with me deeply: organizational structure. That is even harder to fix. The organizational friction and the sheer cost of communication— having to sit through endless meetings and coordinate across departments. If you don't eliminate that, even if AI writes all the code in seconds, everything still gets held up waiting. It's true for both individuals and organizations. It's like the early days of the automobile: Britain had the Red Flag Act, requiring a person to walk in front of a car waving a red flag to prevent accidents. They thought cars were dangerous, so a human had to walk ahead. But that completely nullified the inherent power of the car. Interacting with AI purely via standard ChatGPT is the exact same thing— you become the bottleneck, so a 30% boost is the best you'll ever get. Exactly. Only when you let AI run autonomously do you unlock real leverage. But once AI scales up, the organization itself becomes the person walking in front of the car. Right! You finish building something, and then you're stuck in meetings. In a traditional company, from the moment code is finished to deployment, you need seven or eight approvals, each requiring meetings with groups of people. Just like that, two months evaporate.

That brings me to the second point: the mindset issue you brought up. I've realized more and more that most people operate with a territorial mindset: "Hey, that's my org. Don't touch it—that's my turf." "That's my livelihood. That's my mortgage." Exactly! But when you introduce AI, it completely blurs those rigid boundaries, because so much can now be automated. Trying to defend arbitrary turf lines becomes utterly exhausting.

In fact, beyond that article, I posted another piece in our community predicting that by 2026, more and more companies will actually cut half their headcount first as a prerequisite for AI transformation. When you have too many people and not enough to do, you simply can't adopt AI effectively. Exactly. Only when you have a lean team buried under massive workloads will people genuinely embrace AI. Spot on. It might not sound like the gentlest approach, but sometimes it's just human nature and reality. Absolutely. You see it firsthand— for instance, the AI team I'm currently building is very lean. I genuinely don't care who builds the AI, as long as you build it. All I ask is that we adhere to consistent standards rather than adopting random tools willy-nilly, and I'm happy. I don't have any concept of territory. But when leaders manage their own silos, the moment you touch anything in their domain, they claim ownership: "That's mine." And they'll never admit it's turf protection; they'll concoct other justifications: "This isn't built well," "You're harming the customer experience," or "This isn't secure." They'll lean on all those excuses. That's fascinating— cutting headcount specifically to break that barrier.

Could you tell us a bit more about the roles you're currently hiring for— how many headcounts you have and what profiles you're looking for?

We are currently recruiting across three main pillars:

1. Platform and Tooling — We primarily build on the Google stack: Gemini Enterprise, the Gemini model family, and tooling like Antigravity built around it. For the platform side, we are hiring engineers with strong mindsets— that's the most critical qualification.
2. Engineering and Delivery — These are project execution roles. On this side, we operate as human-and-agent hybrid teams. From day one, you are expected to build with AI agents rather than relying purely on manual human labor. We're hiring across multiple levels— from Engineer to Director, to Principal Engineer, Senior Director, and all the way up to VP.
3. AI Lab — A space I'm personally very passionate about. I set up a dedicated team called the AI Lab; it's a very small, agile research team.

Just two people, doing rapid prototyping. The second one is a rotation program, recruiting college students. It's not an intern role; it's long-term, semester by semester, coming in to work alongside us. We can actually learn a lot from these college students. Absolutely, right, they don't need to unlearn anything. Exactly. They don't need to unlearn. For you, it brings a daily shock: "Wait, you can do that too?" "Really, that works?!" "Oh, they actually pulled it off, right?" When you don't have to worry about so-called production-ready software, this approach works exceptionally well.

I'm hiring for all three of these areas, so from engineer all the way up to VP of engineering, everyone should apply quickly! Right, I'll post the community link later along with the job descriptions, so everyone can just look at that directly. Alright, thank you! Thank you, Senior, okay! Hope today's content was helpful to everyone, see you next time! See you next time!