IT Grad Reveals What AI Engineers Actually Build in 2026

Career Exploration
27 followers
6 Video Views·Aug 26, 2026
Maddy Zhang went from MIT to Microsoft, Google, and Amazon, and now she's breaking down exactly what it takes to become an AI engineer in 2026.
We covered her honest take on MIT vs the Bay Area, why she solved 400 LeetCodes, the email filter that almost cost her a Microsoft internship, why she thinks chasing money kills your career, and what AI engineers should actually be building right now (hint: not another GPT wrapper).
If you're a student, engineer, or anyone trying to break into AI in 2026, this one's for you.
AI Generated·May contain errors
Was it the power of the email or the power of your hard work?
Hello everyone, today I have the pleasure of being with Maddy, her journey from MIT,
especially pursuing bio to pivoting to computer science, then working at Microsoft, Google
and at Amazon, all three big tech, and now making videos to be in the world of AI,
how to become an AI engineer. At MIT we have this saying, it's like going to MIT is like
drinking from the fire hose, I totally felt that, not only was I balancing classes,
also job interviews, also social life or lack of a social life, also a lot of extracurriculars,
so I think that was the hardest. Don't try to be an AI engineer because you like money,
I think in general, I'll say not just for AI engineering, I think wanting to do something
just because of the money is a really good way to not have fun and to hate your life, don't do that.
Honestly, I just don't, there's no good shortcut, I mean, yes, you can use AI,
you can use whatever to kind of consolidate the information and do it fast, but I feel like the
best knowledge comes from actually just sitting down and watching it normally without your open
cloud, without your AI agent, and just going through it step by step. So today it's going
to be a guide on MIT plus how to be an AI engineer in 2026, so hi Maddy, can you please introduce
yourself and how did you grow up here and then your journey to MIT? Yeah, thank you for the
intro, so happy to be on the channel. Hi, I'm Maddy, like Hanoor said, I'm a software engineer,
previously worked at Google, did internships at IBM, Amazon, Morgan Stanley, Microsoft. Yeah,
I went to MIT, graduated with my bachelor's in computer science engineering, and I grew up in
the Bay Area, so like around an hour, 40 minutes drive away from here in lovely San Francisco.
As an intro, MIT has been the most popular video on my channel. Most people watching here,
either engineers or getting into engineering, are passionate about MIT, so I want to know how
MIT has evolved. So back in 2016 to 2024, everyone was really passionate about getting into MIT to
pursue computer science and research, but now more people are passionate about leaving college to do
startup. It's a new wave we are catching, then to be in college. What are your opinions, how MIT,
or like the dream to be in college, has evolved? Because right now, if I go to UC Berkeley,
Stanford, no one is thinking about staying in college. They're not even thinking of graduating,
they just want to get into the startup ecosystem. Yeah, that's a great question. I'm biased. I
graduated a semester early, but I did graduate with my bachelor's in computer science. I would
say, unless you have like a really, really, like, unless you really, really know what startup you
want to do, I would still say err on the side of staying in college. College isn't just about,
you know, the grades, about the diploma. I personally really liked MIT, not just because
my professors were amazing and I had hard classes and liked the classes, but also because of the
opportunities, both implicit and explicit. So what I mean by that is, because I was with a group of
people who were really smart, smarter than me, very passionate, I was able to get inspired by
them. And also if they applied to something, I would know that I could apply to that. If they
got an internship here, I'd be like, oh, okay, like this, this company accepts interns, I should
apply next year. So maybe I, maybe my viewpoint is a bit outdated because I did graduate before 2024,
unfortunately. However, I still think that it's amazing if you know exactly what you want to do
it, especially with AI nowadays, it's so easy to create a product, to get marketing, to do all that.
But if you're not sure, I would still highly recommend staying in college, unless you are
very sure that that's what you want to do. Interesting. Because like, I think that's a
mindset that everyone pretty much had to 2024 because like you grew up in Bay Area and I want
to know, first of all, like what college options you were considering in Bay Area versus MIT?
Because a lot of people who grew up in Bay Area kind of like want to go to MIT because they want
to get out of this community. They're bored with the most competitive area in the world.
So they feel like, you know, they have struggled a lot. Now they want to go out. What was your
reasoning behind going to MIT and leaving the Bay Area? Yeah, great question. I love the Bay.
I love my family in the Bay. However, like you mentioned, I did want to see what it'd be like
living outside the Bay. When I was deciding what college to go to, I didn't really know exactly
what major I wanted to be. The only thing I knew is I definitely wanted to be either in science or
math or technology of some sort. So as a result, I was targeting specifically colleges that were
very good in that. There's a lot of them out there. So engineering and math. Yeah, of course.
No college better than MIT in the world. I also didn't know too much about MIT growing up.
Like I knew it was a school that people went to. However, I will say a lot of the people in my
science, like I did competitive math and science competitions. A lot of my friends from those
circles ended up at MIT. And then the older ones, when they would go back and visit us and talk
about like how it was like, I realized that it wasn't just like it's a prestigious school with
good classes. They really seem like they found their people were learning a lot, were really
passionate and also didn't, you know, get frozen to death in the Boston winter. So as a result,
as I grew older in high school, I decided, OK, this seems like a stretch goal for me,
but it seems like a school that I really, really fit in. And then once I got there,
I realized how much how good of a fit it was for me. I will say for me personally,
I think it was the best fit that I could have ever imagined in a university. So very happy I went.
Home is a place where you feel yourself like you wanted to feel yourself in the world of
mathematics and, you know, engineering. And that was your dream. So like you could feel yourself
pretty much at MIT because the competitiveness was really high, pretty much like when you were
considering schools. What are the schools you got accepted to, if you had to mention?
Yeah. So I got accepted into like a couple of other ones that I was really considering,
including UC Berkeley, like Princeton, Yale. Those were the ones that I was highly considering.
All of them do have great engineering departments and computer science departments.
Something that I really appreciated was all of these schools have kind of a visiting weekend
where you if you get admitted, you can go visit. I really had fun in all of them. But the one where
I truly, like you mentioned, felt at home was MIT. So I was yeah, it was it was honestly such
a blast. That particular visiting weekend, MIT, it's not just the administration hosting events,
students host events. So I remember like playing like life-size Hungry Hungry Hippos,
making nitrogen ice cream, playing the bubble soccer, just so many fun things hosted by students
who also were super busy with problem sets and stuff. So I knew like, oh, this is a place where
people are passionate about stuff, not just because it's for a grade. They're just passionate
people. 100 percent. And what is a typical dream of an MIT student? Because like when I go to any
college here, like right now we are in a wave that every college student wants to do a startup.
What would you say is different about MIT? Do you think people are like,
you know, also catching the wave of startups? Because like there are good and bad people
coming out of every college. But MIT has really held the prestige in Silicon Valley because a lot
of the great founders such as Cursor Founders, they are from MIT. A lot of people like Cognizant
are from MIT. So like all these tech tools which require tough engineering are from MIT. But at the
same time, like I will not mention Delve. I think let's not talk about it. We don't talk about Delve
or SVFs, Liz, thank you. But in terms of like what kind of what kind of like wave you thought
MIT students were catching, what's the biggest dream of MIT students? I think broadly it's just
to make an impact in the world, hopefully good impact in the world. So in like, you know, nowadays
I think the fastest way to make impact is to join like startups or to make your own startup.
I will say though that like a lot of my friends, like all they want to do is go into academia and
teach. And I think that's really admirable. And that's another way that you can really make an
impact in the world because it's not just you fulfilling your own dreams. You're also inspiring
the younger generation to follow their dreams and to get into math and science. And I think,
to be honest, like I feel like people do not value math and science and that kind of those kind of
academics as much as I hope they will in general. So yeah, I think it's broadly just making a good
impact in the world, however that may be. Some people find that doing startups, some people find
that founding startups, some people go into academia, some people, you know, work in big tech
because, you know, even if you are a small cog in a big machine, you can still launch features that
hit billions of users. That's what makes MIT different because they are so passionate. All
of my friends, if I want to mention my friends at MIT who I've interviewed in the last eight years
of my YouTube career, all of them have big passion for research. So a lot of them have like gone into
like Jane Street Citadel, like they have caught the big tech wave where like, you know, so I want
to give some summary. So like in 2016 to 2020, everyone wanted to become a software developer
at Google, Microsoft, Amazon, etc. or Netflix. 2020 to 2022, everyone wanted to get a remote job.
2020 to 2022 to 2024, people wanted to get into Jane Street Citadel because they pay the highest,
you know, core finance salaries. Now everyone either want to work for an AI company like XAI
OpenAI or a startup or want to create their own startup. But what makes MIT different is what I've
seen is people don't catch waves at all. People always pursue what they're passionate for and
they try everything. So my friends, they would work at Citadel, Jane Street, think if it's okay or not,
and then move on to what they're really passionate towards, such as research, working at a
research lab at MIT. And if that works out, then they pursue and do PhD. Some of my friends who are
from IIT to MIT, they went to do some research labs and now they're doing PhD in Silicon Valley.
So that's what I think makes MIT unique. Any stories you have about some passionate folks there?
Yeah, I, for example, I had friends, you mentioned all the quant trading firms who did like Citadel
internship, Jane Street made a lot of money, but then decided to just do a PhD and are very happy
doing that. I have friends who founded startups and now are like billionaires because of the
startups, like for example, like as an example, like MIT startups that recently did that is like
I have friends who were like among the earliest founding members, who were among the first
engineers in startups like OpenAI and Anthropic and are doing very, very well. So I think it all
depends on what you're passionate about. I a hundred percent agree. I feel like in my opinion,
I don't think a lot of MIT students like do things just because of money or because of prestige. I
think they do things because of what they're passionate about. Again, like I said, many of
my friends, not to knock on quant trading, many of my friends who are really good at math, like
did make a lot of money doing trading, but decided that wasn't what they're passionate about. That
like, wasn't what they want to do with their life and then pivoted. So I think it just like,
I feel like a lot of my friends have the ability to do whatever they want. So they try a lot of
things, decide what they're really, really happy doing and then stick with it. Awesome. Now I want
to talk about how to make best use of college and especially MIT, because you have worked at IBM,
Morgan Stanley, Google, Amazon, Microsoft. Was it the power of the email or the power of your
hard work? Great question. Sorry, power of email? What do you mean? So like MIT.edu, is it like
really powerful enough that every recruiter will be like, oh, Maddie, when can we get on a call?
I see. I will say as a first and second year, I found it actually quite hard to get emailed back.
Really? As an MIT student? I think it's like, I was also an MIT student without that much experience
on my resume, but once I got like those first one and two internships, I think like things really
picked up. So I will say, I think having the name helps, but it's not like the be all end all. You
also still have to pass the technical interviews and do leet code and you know, all of that fun
stuff that everyone loves, right? But yeah, I think that it was, I think the advantage for MIT
is yes, like the name probably does help, but having, I think having experience can make up
for it. But I think it's also knowing what programs are out there. So for example, I only
need to apply to some programs because my friends either were applying and wanted to study together
or like I knew they applied the previous year and then applied there. So I think that's a
really strong motivator as an example. So like internship programs. Yeah. As an example,
like I have friends who are really smarter than me, really smart, did all like wonderful companies,
but in college, like they didn't really know like what internship programs are out there. So they
didn't even try. So, you know, like it's cliche, but you miss a hundred percent of the shots you
don't take like at a place like MIT or other colleges where there's a lot of just smart
people who are trying their best and grinding, you hear of more things and more opportunities.
A funny instance that happened, I can laugh about it now. Apparently my MIT email filter
slash Gmail filters filter out Microsoft recruiting emails. Really? Into spam. I learned
this because I remember like I'd gotten like a, I intern at Microsoft and I was like, Oh,
like, I wonder why they're not reaching out to me for the summer internship. Cause I did
a wintership. So I like talked to the recruiter and she was like, Oh wait, no, we like reached
out to you. Check your inbox. I checked, I saw nothing. And she was like, what? Like,
we definitely sent something to you. So I checked in my spam. There's like, Oh, you know,
literally an email saying, because you did an internship already, we're going to fast track
you to the final round. Uh, you know, yada, yada, yada. I never, I never saw that email
because it was in spam. Oh my God. I don't know if it was, you know, MIT filters. Cause I filter
MIT email into my Gmail or if it's Gmail filters, like, you know, they don't want good engineers
working for Microsoft conspiracy theory. Um, but that was an example of how I missed almost missed
opportunity because of that. So if you have to say one thing that really helped at MIT to grab
all these opportunities, what would, would be, was it the friends? Was it like your friend circle?
If they were applying for those programs, you applied and that's how you ended up into those
internships. Or would you say that you had to be out there as much as possible,
like a normal college student to get those opportunities?
Yeah. Um, I think primarily you still have to grind and go to the, all the job fairs. I
went to so, so, so many job fairs throughout my time. Yeah. It's also, to be honest, like,
I don't think the large companies like hire as much at job fairs. So I did a lot of job fairs
where I wait so long in line only to be told like, Hey, apply online. Did a lot of online
applications. Um, I think, so I think there's two parts. I think, um, in terms of getting the
internship, it really helped to like, just like apply online and do all that in terms of passing
interviews, definitely studying with friends, like doing LeetCode with friends, whiteboarding,
rubber ducking with friends was really, really useful.
So how many LeetCode problems do you solve if you're comfortable sharing? Because I think
a person outside MIT has a bias that an MIT student needs to have rest rigor. That's what
exists outside. That's I would say. Um, I solved around like 400-ish LeetCode.
Wow. I've solved a hundred by the way.
I am not naturally good at algorithms. So it takes me a while to get into the,
so you have to do the grind.
Interesting. So now I think I want to go a bit deeper. So I think you said you still had to go
to all the career fairs, like a normal college student. You still had to apply for all the
programs that your friends were applying to get internships and full-time jobs.
What would you say had been the biggest struggle at MIT for you?
Um, I think definitely balancing everything. Um, at MIT, we have this saying, it's like going
to MIT is like drinking from the fire hose. Like I totally felt that. Not only was I balancing
classes, also job interviews, also social life or lack of a social life. Um, also,
like a lot of extracurriculars. So I think that was the hardest because, um, there were just so
many things that you could be doing. And I saw all, a lot of my friends doing like literally
everything. Um, and it was hard sometimes to like, be like, Oh, you know, like I'm comparing myself
to like all my, a lot of my friends who like have like X, Y, C internships, like, you know,
lots of clubs, like great 5.0 G or like we have GPA out of 5.0, some like full GPA, everything.
Um, so just like sometimes I think it's knowing how to balance your time is very,
very important and something that I didn't really learn until the third and fourth year.
And if I were to go back, I would definitely have tried to do better my first and second years.
A hundred percent. And now I think outside college, you have having a better time,
you would say, right? Of course. I think it's a dream of course, to be in big tech. So to learn
how you figured out your passion, I want to know from your internship. So 2016, you went to MIT,
you graduated in three and a half years. So 2017, you got a first internship. Where was it?
Um, it was at a very small startup. I was doing mobile dev for them. Um,
they're doing like a gamification platform where you can, uh, get nonprofits to sign up and then
people can, uh, complete challenges and donate to the profit nonprofits for prizes. Uh, so that was,
it was really fun, but I figured after that, I don't think like mobile dev was what I was truly
passionate about. So then, uh, the next summer I interned. 2018 summer. 2018 summer, I interned at
a mid-sized business analytics company. Um, did a lot of like D3 stuff, a lot of visualizations.
Um, I really liked that. Um, so I really liked that and decided, oh, like I think front end is
fun. Full stack is fun. Um, I, then I did 2018 winter. I did an internship at Microsoft. Um,
I was working on the Xamarin team. Xamarin? Uh, sorry. I was working on the team that makes
Visual Studio. Uh, so this was like a lot of like, uh, C sharp and IDE tools. I was like, oh,
like this is, this is fun. It's like a lot of, it's not just working on a product that people
see, but it's also working on dev tools. In the summer of that year, I interned at Morgan Stanley.
So this was, I was curious what it'd be like to work in the financial industry. I was working on
like visualization dashboards for Brexit at the point, because this was the summer during which
Brexit was supposed to happen. I think it happened like a couple months after. Interesting
project. Wow. And so it was fun. And, uh, this is kind of how I started with like some like
regulatory stuff and learned the EU stands for a lot of, uh, regulations and rules.
And then after that, my last year I worked at Microsoft again, same team. Cause I really,
really liked it. Um, I actually would have probably converted onto that team if they
had headcount, but they didn't. After that I did a spring co-op at IBM. At IBM, I basically worked
on a, uh, like a test infrastructure for one of their spreadsheet projects. Uh, that was a lot
of fun. It was also close-ish to MIT. It was in Massachusetts. So I got to hang out with my
friends and work there. The summer I did this weird thing where I actually did a summer internship
after I, I did both the spring co-op and summer internship after I graduated. So summer I interned
at Amazon and I was working on Alexa, working on basically a clustering news coverage of Alexa and
comparing it with competitors out there to see like what places Alexa news was weakened and
where they can improve. So that was my internship experience. And that was the first time I used
Python in production. And I was like, Oh, I actually do like using Python besides doing
the code problems, which then brings us to Google. My first full-time job. I was at Google for over
four and a half years. Wow. Worked in Google health and then Google search. Full stack. Yeah.
Full stack. So, uh, yeah, like I think I mentioned to you that I'm very, very passionate about health,
almost studied biology at MIT. Um, really wanted to mix my passion in biology and computer science.
And then later I switched to search ads because I just, uh, I think Google, it's not uncommon to
switch teams after two or three years. So I figured it'd be something to, I figured it'd be nice to try something new.
So that's kind of how my, my journey was. That is really interesting. And what was, what would you
say will be your inflection point that made you figure out, okay, full stack is something I'm good
at and I really enjoy as well. Um, I guess for the inflection point, I think it was after I got
promo at Google, I was like, Oh, you know, like I can actually do this job well enough that people
are willing to like, give me, get me to the next level. Awesome. So now I want to know about like,
in this wave of becoming an AI engineer, what would you say is the right trajectory to follow?
I would say first and foremost, uh, don't try to be an engineer because you like money. Like I think
in general, I'll say not just for AI engineering. I think wanting to do something just because of
the money is a really good way to not have fun and to hate your life. Don't do that.
I would say definitely learn the basics, like understand, like, you know, how to like what,
how evals work, what, what harnesses, um, like MCP, prompt engineering, et cetera. A lot of different
companies and courses out there give you good basics, even if you don't officially study AI
engineering. Um, if you're in school, you can definitely take classes. So I guess my, my answer
right now is more geared toward people who are already out of school. Um, and I think nothing
beats building stuff and just like seeing if that's something you like. So for example,
you can like build your own personal rag or you can like, uh, build your own MCP server and see
if that's something that you find fulfillment and you find fun also. Yeah. Building things makes it
so you can put things on your resume. And if you are, even if you don't have official, uh, AI, um,
schooling, then sometimes having things on your resume can make it much more likely for a company
to actually realize like, Oh, you know, you might have the skills that we're looking for.
But I will say first, like, I will say regardless, nothing beats having good fundamentals. So whether
it's engineering, whether it's anything else, like you definitely don't want to just like rush
into things, like definitely take the time to understand the fundamentals. Cause I can assure
you that basically having the fundamentals is going to save you a lot of time in the future,
even if you don't think so at the time. Interesting. So here comes the most important
question. Like a lot of people like build like a small chat, GPT wrapper, call that a startup here,
raise funding and say as if they're building something revolutional. So in terms of the
projects that really excite AI engineers, and especially in this job market, what would you
say is something? Yeah. Um, I would say it's like probably less about like how, uh, if you build a
map for stuff, I think it's like some of the stuff that's more like foundational and building blocks.
So I would say like if you build MCP or if you build a rag or if you build like a harness system,
that's like definitely more impressive to recruiters and to AI engineers than, um, you know,
doing like a wrapper and like trying to like market it as a full product. Awesome. Now the big
question, do you think Silicon Valley is overhyped? I genuinely don't think so. I know I'm biased
because I live here, but I feel like there's so many smart people, cool things to do, cool startups
and just like amazing events. And it's really, really worth it to be here. Considering your path
from MIT to Silicon Valley, you left Silicon Valley for MIT and then you left MIT for Silicon
Valley. What would you do differently if you had to like restart your journey from today to get into
tech industry? I, to be honest, would get into coding earlier. I only started coding last year
of high school. And when I started at MIT, a lot of my friends have been coding while they were
seven, eight, you know, from very, very early on. Something else I do differently is I don't
usually ask for referrals when I try to leave my company. And because I'm kind of just like
afraid to admit to me and my friends that I actually want to leave. Hopefully this doesn't
happen, but if I ever need to leave again, I definitely will ask for referrals because I
think it's like makes it much more likely that my resume will be seen by companies out there.
See, I actually like wanted to do a challenge that I, I wanted to get hired without referrals,
but it actually costed me an extra month to get a job. I, if I had to do from scratch as well,
I'll do the same because I think referrals make a big difference inside a community.
And last question would be, so, so you started coding in the last year. So what would you say
the one skill you had that helped you get into MIT? I think I'm a very good generalist. I'm not
like, you know, I'm a level math, but I will say I'll list out some of the stuff I did. So for
example, I was president of like four different clubs in high school. I did varsity track,
varsity tennis, did student, uh, did undergraduate research at UC Berkeley. Um, I also competed in,
in the science competitions, like did, uh, made the semifinals of USABO. Also do, uh, also was
founded club, like founded biology club. And like we did a bunch of outreach as well. Um,
and also was, I was also a valedictorian. So I had a pretty good grade. Sorry. I was also
valedictorian. So I had the best grades at my school. Uh, that being said, that being said,
my, my strength was not computer science specifically. So when I applied to MIT,
I did not declare computer science. I think that's fine. I think MIT saw that I was just
passionate about learning and I would figure out my path when I got there. And that's what I did.
Awesome. Everyone talks about building an AI engineer. Even you mentioned in your videos
that everyone needs to learn the fundamentals, like vector databases, like, uh, learning machine
learning, fundamentals, algorithms, watching Andrey Karpathy videos, watching like the videos
from Andrew NG, like from his course on course setup, what would you add? Like how to like do
something? Is there a shortcut to becoming one? You would you say like, because a lot of people
get lazy. They don't finish the courses. They know everything, what they want to do. They know
there are like two to three hour videos and they add up to 30 hours. They don't want to finish.
Is there a shortcut? Honestly, I just don't, there's no good shortcut. I mean, yes, you can
use like AI, you can use whatever to kind of consolidate the information and like do it fast.
But I feel like the best knowledge comes from actually just sitting down and like watching it
normally without your open cloud, without your AI agent and just like going through it step-by-step.
Because I think there's a new AI wave happening that is anti-AI. I have actually like stopped
using AI for emails. I write my own emails. Stopped using AI for so many things that brings
back my intelligence. And same has to do in learning as well. People have to learn like
it's 2015 and work like it's 2026. Yeah. Awesome. So that'll be it. Thank you so much Maddy for
sharing incredible insights to your MIT journey and also the guide to be an AI engineer in 2026.
And yeah, and check out the collab that we're going to do on her channel as well
with the link in description below. Awesome. Thank you so much. Thank you for having me.
