why this exists
i have a theory that every engineer who builds a personal site eventually writes a "why i built this" post. it's almost a rite of passage. a ritual performed by people who spend their lives telling computers what to do and then, for some reason, feel the need to also tell strangers about it. so here's mine. i'll try not to be boring, but i make no guarantees.
the short version
i needed a place to put thoughts that don't fit in a commit message.
the longer version
i grew up in joão pessoa. it's a beach city in northeast brazil where nothing moves fast, the sunset lasts forever, and your biggest decision on any given tuesday is which beach to go to. i spent my teenage years teaching violin to kids who couldn't afford lessons, got a technical degree in violin, then another one in piano, then started guitar, then realized i cannot sing to save my life. i was also a semi-professional swimmer. and before you ask, yes, i did try a new sport or hobby basically every week. some stuck. most didn't. surfing is the latest attempt. the ocean is winning.
but the thing that actually stuck? minecraft.
i'm not kidding. that's the origin story. i wanted to build a modded server when i was a kid, and that's how i learned my first lines of code. from there it was java, then python, then "wait, you can make machines learn things?" and suddenly i'm reading papers at 3am about gradient descent like a completely normal teenager.
at some point the coding stopped being a hobby and became the thing i do instead of sleeping, eating, and maintaining relationships. i'd pick up a problem, any problem, and disappear into it until it was solved or i physically couldn't stay awake. that's still how i work. it's not healthy and i don't recommend it but i also haven't figured out how to stop.
when i'm not doing that, i'm probably at home playing video games. elden ring is the greatest game ever made and i will not be taking questions on this. expedition 33 just dropped and i haven't been productive since. if fromsoft ever goes public i'm putting my life savings in.
university (sort of)
freshman year at ufpb i joined TAIL, our AI research league, and started doing the things you do when you're 18 and think you understand machine learning: recommendation systems, NLP, trying to predict crypto with chaos theory. it didn't work, obviously, but the math was beautiful and i was too young to know i was wasting my time, which is honestly the best state to be in.
i should mention that i've gone to maybe one class per semester since i started university. i basically show up for exams and leave. the brazilian education system and i have an understanding: i don't waste their time, they don't waste mine. somehow this has worked out so far. i do not advise this strategy, but i also won't pretend i regret it.
then reality called.
someone asked me to build a system that could grade essays for brazil's national exam. the one that determines whether millions of students get into college. i said yes because i didn't know enough to say no. two weeks later we had an MVP. it became the first app in the country to do it. a million essays. featured on national television. i was 20 and deeply confused about how that happened.
that same year i wrote golang systems that processed 40% of a state's fiscal invoices. built ML models that made 400 hires unnecessary at a government agency. got into uber's engineering fellowship (they take fewer than 2 in 100). co-founded a software house and shut it down when i realized i wanted to build products, not services. a lesson that cost me nine months and saved me nine years.
all of it in about 24 months. looking back, none of it felt planned. it felt like falling forward and hoping the ground would catch up.
11x
here's a fun one. 11x was a series B startup that wasn't hiring anyone who wasn't a CTO, a founder, or very very senior. i was none of those things. i was a kid from brazil who really wanted to work in silicon valley.
a friend connected me with one of the engineers there. he's the one who opened the door and put me in front of everyone else. from there, i basically just showed up with work. didn't pitch myself, didn't sell a vision. just said give me two weeks and let the code talk. they did. they extended me a contract.
that engineer became one of my best friends. he's from stanford, and watching how he thought, how everyone there thought, so far outside the box that the box didn't even exist anymore. it hit me. i needed to get out of my town. i needed to be around people like that.
i became their only remote worker. for a long time, the only one. but i shipped, so nobody complained. they gave me a product to build from scratch. i built it. they gave me another one. i built that too.
so i started researching. how do you get to stanford when you're already halfway through your undergrad? can't transfer. don't want a master's. the only option was visiting student, a program that's extremely competitive. i applied. got in. 100% scholarship on tuition. only brazilian in my cohort.
first thing i did? enrolled in CS229. andrew ng's machine learning course. because obviously i thought i was smart enough for that.
i was not.
i couldn't even finish p-set 0. the math at stanford is not the math i learned in joão pessoa. it's not even the same sport. i had never in my life thought i wouldn't pass an exam. CS229 made me think that every single week. before the final, i studied for three days straight. not three days of studying. three days of not sleeping. 72 hours. personal record. i passed, but my ego didn't.
san francisco
11x brought me to the bay. stanford humbled me. and sf... sf is complicated.
here's what nobody tells you before you move here: san francisco is the most depressing beautiful city you'll ever live in. the weather can't decide if it's summer or winter, sometimes within the same block. you'll walk past a $4,000/month studio apartment and step over a needle on the same sidewalk. half the city looks like a zombie movie set. the tenderloin exists ten minutes from where billionaires eat $40 salads. it's a city that runs on cognitive dissonance.
and yet.
the energy is undeniable. in joão pessoa, ambition is something you keep quiet about. you don't tell people you want to build a company, because they'll ask why you don't just get a government job. in sf, "what are you building?" is the local "how are you?" and everyone actually means it. the barista is building a startup. your uber driver has a pitch deck. it's exhausting and inspiring in equal measure.
sf is not a place you love. it's a place you tolerate because it's the only city on earth where this specific flavor of insanity is considered normal. you live in a shoebox, eat overpriced food, dodge the occasional person screaming at a fire hydrant, and somehow feel like you're exactly where you need to be. it doesn't make sense. that's kind of the point.
after 11x, the math was simple. i'd already built two products from scratch inside a series B. now i wanted to build something from scratch that was mine. not inside someone else's company. my own.
that's when i joined sable as founding team. sequoia-backed, building from zero. the kind of early where there are no processes, no playbooks, just problems and a whiteboard. the kind of early where everything is on fire and nobody panics because there's no time.
so why does this blog exist
honestly? i'm not sure yet.
partly because writing forces you to think clearly, and god knows i could use more of that. partly because i want a record. ten years from now i want to read this and cringe at how little i understood about what i was getting into.
partly because the best founders i've studied, the ones who actually built lasting things, all had some version of this. not a polished brand. not a ghost-written thought leadership blog optimized for linkedin engagement. just a place where they worked through ideas in public, before they had any credibility to do so.
i'll write about engineering, about what it's like being brazilian in a city that thinks açaí bowls are a personality trait, about building a company when you're young enough that failure is basically free. maybe about music. maybe about why elden ring's open world design is better than most software architecture i've seen. maybe about how sf somehow convinced an entire generation that paying $7 for coffee is reasonable.
no schedule. no promises. just whenever something feels worth saying.
let's see what happens.