Get live statistics and analysis of Michael Guan's profile on X / Twitter

we’re hiring at @finalroundai. this isn’t a normal job. it’s a chance to help us replace interviews with proof. with over one million candidates from 100+ countries already securing their dream jobs, we're transforming careers and breaking new ground. open roles: > founding data engineer → design pipelines & turn raw signals into predictive insights > founding fullstack engineer → ship core product features > founding ai engineer → architect llm/agent systems that evaluate real work in real time our culture: > growth mindset — we’re always learning, breaking walls, and finding new paths > audacious & bold — we aim to set trends, not follow them > speed & scrappy — move fast, stay focused, punch above our weight > ownership — self-driven, outcome-oriented, no excuses > team player — diverse collaboration, collective excellence you are: - technical & product-obsessed — you care about how things feel as much as how they run - low ego, high ownership — there will be unsexy work, you’ll do it anyway - ambitious, fast-learning, relentless about standards - excited by ambiguity, there’s no playbook here, you’ll be the one writing it we don’t index on years of experience. we index on energy, values, and slope. comp + equity will reflect the impact we expect. if you’ve ever wanted to help redesign hiring from the ground up, this is it. 📩 DM me or email hr@finalroundai.com (referrals welcome).

happy diwali 🪔 to everyone celebrating, may your week be filled with light, energy, and small wins that compound. our team at Final Round AI and Work Trial AI is spread across india, us, and beyond, and today the chats are full of photos, sweets, and laughter. building globally also means celebrating globally. here’s to brighter ideas, stronger teams, and new beginnings. #Diwali2025

in 24 months, we went from a hacked-together prototype to $10M+ ARR and 7M users, w/o doing a single “traditional” interview. the spark: after selling my last startup i wanted to build something ai-native, not just “chatbot #243.” i kept thinking about jarvis from iron man, an ai that doesn’t wait for you to type, it proactively helps you in real time. our first product was meeting copilot. it listened in zoom calls and suggested prompts. one user asked: “can it help me in an interview 10 mins from now?” i said yes, $99/mo. he paid on the spot. that was the moment. the pivot: interviews are a huge pain. people spend months searching and prepping, but lose chances because of nerves or blind spots. so we built interview copilot: a real-time teleprompter for interviews. people called it cheating. we called it the future. one user had been unemployed for 8 months. with us, he landed a job in 17 days. hitting $1m arr: an advisor told us: “want more revenue? raise your price.” so we went from $99 → $150. people still paid. overnight, revenue jumped 50%. then we joined @hf0. 12 weeks of growth bootcamp. shut off free trials - shipped 4–5k seo pages a week - ran 1000+ tiny experiments - arr: $1m → $3m in 3 months. scaling to $10m arr: no silver bullets. just compounding small wins. we built an internal ai-native a/b test platform so anyone (eng, ops, marketing) could run experiments. good ideas came from everywhere. we turned influencer marketing into relationships , inviting creators to live with us, not just one-off ads. we threw parties where 800+ founders, investors, devs showed up. culture became growth. (dm me if you wanna come to our party, we are still hosting!) what’s next: today final round ai has 7m+ users and $10m+ arr. but arr isn’t the mission. we’ve already stopped doing interviews inside our own team. instead, candidates come in and actually work with us for a day or a week. now we’re productizing that into something new: ai agents simulate coworkers, managers, even customers. candidates log in, do real tasks, and we generate signals about how they think and collaborate. imagine replacing interviews with actual work. lessons learned: - controversy is a good sign - raise prices earlier than you think - growth = 1000 small experiments, not 1 hack - culture compounds just like metrics we’re just getting started. if you believe hiring should be fairer and more real, would love to chat because in the ai era, it’s not “what you know” — it’s how you think and solve problems.

building culture, and keeping it consistent, is hard. for startups, culture isn’t just a nice-to-have. it’s the only thing that keeps people around when things get tough. as a founder, i think about this every day: how do i become the kind of leader my team isn’t afraid to give feedback to? bc real culture isn’t built from values on a wall, it’s built from convos you don’t shut down. #foundermode

less than 24 hours until launch. jay and i were laughing last night, remembering the first prototype we hacked together. a duct-taped “meetingcopilot” that somehow convinced a user to pay $99 for interview prep. fast forward → tomorrow we launch @WorkTrialai. a platform built from the same obsession: interviews don’t show you the truth, real work does. nervous? sure. but mostly grateful to be building this side by side with a co-founder @Doudoujay who’s just as crazy as me. #worktrial #foundermode

we didn’t set out to build an “interview tool.” early testers used final round ai in all kinds of weird ways. but one pattern kept repeating: people practicing for job interviews. that wasn’t our plan. but it was their need. lesson learned: you don’t always find product–market fit. sometimes it finds you. #founders

Most startups don’t die from competition. They die from bad hires. I learned this the hard way while building Final Round AI. In 2 years, we scaled from 0 → 7M users and $10M Annualized Revenue. Along the way, we hired dozens of people. Some were game-changers. Others… cost us months of runway and momentum. It reminded me of the old story: the young warrior who sets out to slay the dragon, only to become the dragon himself. AMA about @WorkTrialai x.com/i/spaces/1rmxP…

after years of grinding… we finally made it to @hf0 w/ my co-founder @Doudoujay. but before final round ai, here’s the rewind ↓ college – flew to brazil on a consulting project. a client asked: “why not start your own company?” → planted the seed. junior year – sneaker hustle, 200 yeezys stuck in shanghai customs during covid → cashflow collapsed. senior year – ngo consulting took me to 100+ countries, saw how early teams really operate. mba – yale → realized few classmates actually wanted to start. dropped out, joined a unicorn. mid-20s – carbon credits startup → too early. private jet travel hack → fun, not scalable. european carbon platform → wrong partners, wrong market. summer ’23 – sold my last startup. gpt-3.5 just dropped. everyone was building chatbots. we wanted jarvis. oct ’23 – shipped “meetingcopilot.” a user paid $99 to use it for an interview. that was the spark. apr ’24 – joined hf0 in san francisco. 12 weeks, 100 experiments. revenue jumped 3x. today: 7m users, $10m+ arr. and the chance to rethink how hiring itself should work. for hf0? it only took: one ticket to the valley, a duct-taped demo, hf0 as our bootcamp, and a decade of obsession with building. #builder #founder

founder life means the week doesn’t end on friday, it just shifts gears. next monday, we launch something we’ve been obsessed with for months: @WorkTrialai. a platform that lets teams see how candidates actually work inside a safe slack, github, notion sandbox. not what they say, not what’s on a resume, but proof. the feeling right now is a mix of nerves + gratitude + quiet excitement. we’re building for every founder who’s lost runway to a bad hire. for every candidate who’s better at working than at interviewing. back to work, monday’s coming fast. #founder #buildinpublic

building final round ai → new product(shh we're gonna launch in a few day) has taught me this: the hardest part of startups isn’t building features. it’s unlearning the habits that no longer serve you. – in the early days, speed matters more than structure. – once you hit scale, discipline matters more than speed. – in hiring, the same lesson: stop relying on old tools (resumes, interviews) once they no longer work. the only constant is knowing when to let go. #founders

Grateful to our team and the millions of candidates using us to land jobs around the world. my favorite user quotes from this week ↓ our users said it best: > “it felt like i had somebody behind me typing everything.” > “the hints are better, i don’t want to sound scripted when i’m talking.” > “if i could put in custom questions, that would have been awesome.” > “the answers you guys gave are too broad sometimes and too long.” > “bullet points are helpful because they’re easier to read.” > “i did four interviews with four companies and i got all four jobs.” > “the active response is really great because i can just glance over and get those little footnotes as it goes.” > “the biggest thing i liked was the confidence boost… the real-time prompts, the structure, it reduced my anxiety.” > “i don’t mind the paragraph, but it’s easier to just glance over at a bullet point and remember… reading word for word is hard.” this is why we keep building. not every feature is perfect, not every answer is short enough yet, but every day, candidates tell us they feel more confident, less anxious, and better prepared to land the job they want. onward. #foundermode #CustomerAppreciation

building Final Round AI sometimes feels like building a dating app(remind me of @hinge's campaign). we want users to delete us: because it means they landed the job, got the offer, moved on. but at the same time, we want to keep them, to be their trusted co-pilot as they grow, switch roles, and chase the next chapter. it’s the strange beauty of purpose-driven products: success means you become less needed. but that’s okay. because when people leave better than they arrived, that’s the real win. #foundermode #reflections

h1-b application fees are jumping from $1k → $100k per applicant. the signal is clear: the US is making it harder to bring talent in. but talent itself is borderless. for founders, that means opportunity. we can hire globally, build globally, and scale faster than ever. @finalroundai has office across India and U.s already. the next great companies won’t be built where visas are cheapest. they’ll be built where talent is strongest. #founders #FutureOfWork #modernwork

we’re hiring at @finalroundai. this isn’t a normal job. it’s a chance to help us replace interviews with proof. with over one million candidates from 100+ countries already securing their dream jobs, we're transforming careers and breaking new ground. open roles: > founding data engineer → design pipelines & turn raw signals into predictive insights > founding fullstack engineer → ship core product features > founding ai engineer → architect llm/agent systems that evaluate real work in real time our culture: > growth mindset — we’re always learning, breaking walls, and finding new paths > audacious & bold — we aim to set trends, not follow them > speed & scrappy — move fast, stay focused, punch above our weight > ownership — self-driven, outcome-oriented, no excuses > team player — diverse collaboration, collective excellence you are: - technical & product-obsessed — you care about how things feel as much as how they run - low ego, high ownership — there will be unsexy work, you’ll do it anyway - ambitious, fast-learning, relentless about standards - excited by ambiguity, there’s no playbook here, you’ll be the one writing it we don’t index on years of experience. we index on energy, values, and slope. comp + equity will reflect the impact we expect. if you’ve ever wanted to help redesign hiring from the ground up, this is it. 📩 DM me or email hr@finalroundai.com (referrals welcome).

in 24 months, we went from a hacked-together prototype to $10M+ ARR and 7M users, w/o doing a single “traditional” interview. the spark: after selling my last startup i wanted to build something ai-native, not just “chatbot #243.” i kept thinking about jarvis from iron man, an ai that doesn’t wait for you to type, it proactively helps you in real time. our first product was meeting copilot. it listened in zoom calls and suggested prompts. one user asked: “can it help me in an interview 10 mins from now?” i said yes, $99/mo. he paid on the spot. that was the moment. the pivot: interviews are a huge pain. people spend months searching and prepping, but lose chances because of nerves or blind spots. so we built interview copilot: a real-time teleprompter for interviews. people called it cheating. we called it the future. one user had been unemployed for 8 months. with us, he landed a job in 17 days. hitting $1m arr: an advisor told us: “want more revenue? raise your price.” so we went from $99 → $150. people still paid. overnight, revenue jumped 50%. then we joined @hf0. 12 weeks of growth bootcamp. shut off free trials - shipped 4–5k seo pages a week - ran 1000+ tiny experiments - arr: $1m → $3m in 3 months. scaling to $10m arr: no silver bullets. just compounding small wins. we built an internal ai-native a/b test platform so anyone (eng, ops, marketing) could run experiments. good ideas came from everywhere. we turned influencer marketing into relationships , inviting creators to live with us, not just one-off ads. we threw parties where 800+ founders, investors, devs showed up. culture became growth. (dm me if you wanna come to our party, we are still hosting!) what’s next: today final round ai has 7m+ users and $10m+ arr. but arr isn’t the mission. we’ve already stopped doing interviews inside our own team. instead, candidates come in and actually work with us for a day or a week. now we’re productizing that into something new: ai agents simulate coworkers, managers, even customers. candidates log in, do real tasks, and we generate signals about how they think and collaborate. imagine replacing interviews with actual work. lessons learned: - controversy is a good sign - raise prices earlier than you think - growth = 1000 small experiments, not 1 hack - culture compounds just like metrics we’re just getting started. if you believe hiring should be fairer and more real, would love to chat because in the ai era, it’s not “what you know” — it’s how you think and solve problems.

less than 24 hours until launch. jay and i were laughing last night, remembering the first prototype we hacked together. a duct-taped “meetingcopilot” that somehow convinced a user to pay $99 for interview prep. fast forward → tomorrow we launch @WorkTrialai. a platform built from the same obsession: interviews don’t show you the truth, real work does. nervous? sure. but mostly grateful to be building this side by side with a co-founder @Doudoujay who’s just as crazy as me. #worktrial #foundermode

we didn’t set out to build an “interview tool.” early testers used final round ai in all kinds of weird ways. but one pattern kept repeating: people practicing for job interviews. that wasn’t our plan. but it was their need. lesson learned: you don’t always find product–market fit. sometimes it finds you. #founders

Emissions platform CYNK announces carbon forwards trade of over 2 mln credits reuters.com/article/climat…

happy diwali 🪔 to everyone celebrating, may your week be filled with light, energy, and small wins that compound. our team at Final Round AI and Work Trial AI is spread across india, us, and beyond, and today the chats are full of photos, sweets, and laughter. building globally also means celebrating globally. here’s to brighter ideas, stronger teams, and new beginnings. #Diwali2025

h1-b application fees are jumping from $1k → $100k per applicant. the signal is clear: the US is making it harder to bring talent in. but talent itself is borderless. for founders, that means opportunity. we can hire globally, build globally, and scale faster than ever. @finalroundai has office across India and U.s already. the next great companies won’t be built where visas are cheapest. they’ll be built where talent is strongest. #founders #FutureOfWork #modernwork

building final round ai → new product(shh we're gonna launch in a few day) has taught me this: the hardest part of startups isn’t building features. it’s unlearning the habits that no longer serve you. – in the early days, speed matters more than structure. – once you hit scale, discipline matters more than speed. – in hiring, the same lesson: stop relying on old tools (resumes, interviews) once they no longer work. the only constant is knowing when to let go. #founders

launching @WorkTrialai in less than 24hrs. we are only just getting started🔮 #worktrial #modernwork

building Final Round AI sometimes feels like building a dating app(remind me of @hinge's campaign). we want users to delete us: because it means they landed the job, got the offer, moved on. but at the same time, we want to keep them, to be their trusted co-pilot as they grow, switch roles, and chase the next chapter. it’s the strange beauty of purpose-driven products: success means you become less needed. but that’s okay. because when people leave better than they arrived, that’s the real win. #foundermode #reflections

this year’s nobel prize in economics hit close to home. mokyr, aghion & howitt were recognized for formalizing what every founder quietly learns the hard way: progress isn’t just about invention. it’s about systems that let new ideas replace old ones. that’s exactly what we’ve been building toward with Final Round AI and now Work Trial AI. for decades, hiring has been running on old infrastructure: résumés, behavioral interviews, personality tests. tools built for a different era, optimizing for predictability, not potential. aghion & howitt called this cycle creative destruction: growth happens when new methods displace the comfortable ones. but incumbents always resist. in hiring, we see resistance looks like: – “we’ve always done it this way.” – legacy HR contracts that make change painful. – the fear that something new might fail before it scales. but progress demands friction. and what we’re seeing now, AI-assisted interviewing, real work trials, data-driven evaluation, is the next wave of that cycle. mokyr said societies advance when they understand why things work, not just that they work. the same applies to hiring. we shouldn’t just ask “can this candidate do the job?” we should ask “how do they think, adapt, and learn?” that’s what Interview Co-Pilot and Work Trial AI are built for: systems that don’t just assess performance, but reveal the thinking behind it. the better they get, the more they teach us, a self-compounding loop where innovation feeds innovation. creative destruction isn’t just an economic theory. it’s the process of building something new while knowing it will break something old. and that’s exactly how lasting progress happens. #NobelPrize #innovation #FutureOfWork #hiring #worktrial

had dinner with a friend at a burmese restaurant last night. the food brought back a summer i spent in myanmar years ago: > teaching english to local students and monks, > helping them set up facebook pages to reach travelers for meditation retreats. it reminded me of a time before “founder” became an identity: when doing good was simple, and feeling useful was enough. funny how those early memories still ground you, esp after a long week of building and chasing goals. sometimes it’s good to remember: before we built things for scale, we just wanted to build things that helped people. (i bet you can't find me from the photo) #oldtime