About
Growth marketer and operator at Scale AI, focused on contributor marketplace growth and…
Services
Activity
3K followers
-
Tuan Hoang reposted thisTuan Hoang reposted thisI’ve tested 3 AEO tools in the last few months: Peec, Profound, and Scrunch. Only one is doing something the others aren't even attempting Scrunch sponsored this post, but I only agreed because what they're doing is actually different. Here's my honest take: With all these tools, you can see how you stack up against competitors for search terms. Basically table stakes in this field The big difference with Scrunch: Measurement + actually serving AI-readable content to bots at the edge This means that humans see your site, while bots see an AI-friendly version built for them Even better, it's optimized by agent visiting your site, so the Googlebot sees your normal site, no risk of cloaking or being penalized. In a zero-click world, the agents themselves are your highest-value visitor They’re going to be making the purchases on behalf of consumers and companies soon One catch: the full bot-serving layer is enterprise-only and needs a CDN setup. The monitoring + audit pieces are open to try right now though If AEO is on your radar, Scrunch is the one I'd put on your shortlist to test out. Link in comments #ScrunchPartner
-
Tuan Hoang shared thisif you're feeling loss and unsure how to start when it comes to AI tooling, this is the best course for marketers on the interwebs right nowTuan Hoang shared thisTo my fellow marketers who are enjoying the Claude life, we're giving away a Mac mini (m4 chip) to power your new AI workflows Here’s how dead simple it is to enter the giveaway: 1/ Comment “🍎” on this post 2/ Sign up on Claude Marketers site (claudemarketers dot com) *Optional* Repost for 10 additional entries Like I said, deaaaad simple to enter I’ll be announcing the winner here on 06/01 at 8AM PDT! P.S. Connect with me so I can DM you if you win P.P.S. This is our way of celebrating hitting 10k users on the site Good luck!!
-
Tuan Hoang shared thisGreat opportunity to work with one of the best teams in the space!Tuan Hoang shared thisWe are hiring! We have two open roles and please apply if you are interested in joining a strong performance marketing team at Uber. Role 1: https://lnkd.in/gRZegHBV Role 2: https://lnkd.in/gtirnTRZ
-
Tuan Hoang shared thisteam TextQL all the wayTuan Hoang shared thisLast week in the Scale AI office, people I'd never met told me TextQL changed how they work. That sentence still doesn't feel real to me. Hear it straight from the Head of Analytics Heqing Huang how 4 business teams at Scale are running 1900 data requests weekly through TextQL Ana. Full case study in the comments.
-
Tuan Hoang reposted thisTuan Hoang reposted thisAI agents are getting put to work on real tasks. Training them to do that well is harder than it looks. For the past year at Scale AI, we've been building environments where agents can practice real workflows - simulated worlds that mirror real software, real processes, and real decisions. They can fail, try again, and get better before they touch production. This process is how these agents eventually become reliable for the world’s most important decisions. Today we're officially launching Scale RL Environments. Nearly half of our new training projects already run through them. We're building more this year. See how we’re enabling the next generation of agents: https://lnkd.in/gHtpwWJHThe Next Frontier of Data Training: RL EnvironmentsThe Next Frontier of Data Training: RL Environments
-
Tuan Hoang reposted thisTuan Hoang reposted thisAfter an incredible 2025 of enabling frontier AI development, we're looking to expand our team with exceptional operators who thrive at the intersection of strategy and execution. You'll be responsible for managing the supply-side of our GenAI business, working cross-functionally to own core processes that directly impact our $1B+ ARR at Scale. You're a great fit if you: - Have driven marketplace initiatives at high-velocity startups or tech companies - Make data-driven decisions and excel at translating analytics into action - Build strong internal/external partnerships while operationalizing complex programs - Can translate ambitious vision into measurable results - Get energized by working on problems that directly impact AI advancement If you're interested or know someone who might be a great fit lets connect!
-
Tuan Hoang reposted thisTuan Hoang reposted thisThis week, Humanity's Last Exam was published in Nature Magazine. https://lnkd.in/ewJTxzsh Created by Scale AI with the Center for AI Safety, HLE is considered to be the most difficult test ever devised, written by thousands of field representatives in maths, chemistry, bio, physics, CS, and humanities with unsearchable precision (now multi-modal). HLE proved one important thing to the industry - the power of originated intelligence. People that chug away in the background to push their corner of the field forward had replication and integrity as their ultimate confidence measure, which the most fluent models structurally failed to replicate themselves. It's so exciting to see these people brought more and more into the research fold of genAI, but there is still a long way to go. This year with agentic workflows and new training environments springing up left and right, there is a silent divide emerging in how we use / capture intelligence to train on. TLDR, real agency will come from the habits of our best truth-seekers. There are many strong beliefs here. Bets that pure enhancements to multi-system workflows or reasoning traces etc will solve for this amount of taste. Or that if we get specialized learning right and crib edge cases with synthetic data, we’ll breeze right past top human intuition. Understandably, there are too many cooks in the kitchen, and we should want to cut through this to the reality we are all training for. As Anthropic showed in enterprise adoption last year with AI safety, there’s just something economically essential about integrity. HLE is a good tonic to this stuff, and Scale has led wonderfully aligned to these ideas. In Davos, Demis Hassabis reminded that there has always been a scientific definition for AGI - boiled down simply as a system that can exhibit all the cognitive capabilities humans can (“and I mean all”). Models may solve question 6 on the International Mathematical Olympiad and can accurately fold proteins under insane complexity, but we’re better off agreeing that true invention is nowhere near exhibited. HLE is a measure of the intellectual frontier, but it shows that for any system to originate breakthrough, we have to look to those who do. Be it Feynman or Hawking or Ramanujan, some of the most efficient drivers of discovery cared as much about the rote intellectual output as they did about the pursuit of truth and all the integrity baked into it. Not long after HLE launched, we began a sister project in stealth - Human Frontier Collective (HFC). HFC is an assembly of humanity’s field representatives that led HLE performance, ML scholarship, and other Scale breakthroughs, and we’re excited to deepen their involvement in frontier research in 2026. You can read more about HFC below. Please reach out if this is of relevant interest to you. Huge congrats to the team, and time to build :) https://lnkd.in/eKWQHETw
-
Tuan Hoang reposted thisIf you’ve ever looked at a fast-growing company and thought, “I wish I can learn how the big decisions are made” — keep reading 👀 I’m hiring for a Chief of Staff to support Cleo S., our VP of Growth & Operations at Scale AI. Cleo is a rare kind of operator who’s scaled global businesses at Uber (China & EMEA), served as a COO and venture investor, and now brings that same rigor and taste to building world-class ops. She’s sharp, fast, and deeply hands-on, and this role is a front-row seat to how top-tier operators drive growth, retention and execution at a world-class AI-native company. This is a true force-multiplier role: you’ll help set direction, unblock teams, turn messy problems into crisp plans, and keep our most important initiatives moving. You’ll have visibility across the business, real ownership, and a direct hand in shaping how we scale (pun fully intended) 📈. If you or someone in your network is analytically sharp and loves problem-solving, collaborating and is genuinely energized by building systems — DM me or apply directly below 👇 https://lnkd.in/gZ-T5U45
Tuan Hoang replied to a comment
4h
Lotfi Drif you just need to apply for a higher rate limit... talk to your AM
-
Tuan Hoang liked thisTuan Hoang liked thisAnyone wanna grab a coffee or lunch? $200/month Claude Max isn't enough to keep the pipeline going and the usage credits go down by 50% in 3 days. Options: a) Get a second Max plan b) Buy more credits, my agent swarm can burn $50 an hour (still cheaper than humans) c) Enjoy the company of human intelligence Again, anyone in the North Shore area wanna grab a coffee?
-
Tuan Hoang reacted on thisTuan Hoang reacted on thisHad a lot of fun at Scale AI HQ in SF this week! Got to participate in our first Demo Science Fair, where I showcased the Comms Cats (Copy Cat and Clean Cat, who draft messages & pull audience lists), whom I created to do the annoying parts of my job and who now save my team 50+ hours a month! (George Quraishi I'm going on vacation and leaving the cats in charge for the rest of the year ok??) I had a lot of fun making my virtual tri-fold board which took me back to my middle school science fair days. Next time I do this I'm definitely going old school and making one out of cardstock and posterboard. I also got to lead my second Claude Automation Lab, and help non-technical Scaliens build automations for their daily workflows. It's incredibly exciting (and kind of terrifying!) to be living in 2026 where someone who's never taken a computer science class (hi) can be leading hands-on workshops on how to use Claude Code. Huge thank you to the GTM Nexus team Erin Ellsworth Jaideen Cobb Cassie Alter for bringing me out and to everyone who stopped by to say hi to the Cat(s) in the Chat!
-
Tuan Hoang reacted on thisTuan Hoang reacted on thisCan an AI agent diagnose a patient correctly without reading their chart? Should it? Today Scale AI Labs released CliniCARE-Bench: 16 agentic systems, 25 clinical care scenarios, 750 real patient cases. A benchmark built to test whether these systems can do clinical work, not just answer clinical questions. Two findings. Error rates on the underlying models reached 34.7%. And when we required systems to be both correct and free of incorrect shortcuts, scores dropped by as much as 14.8%. The second number matters. Agents were frequently landing on the right answer without working through the longitudinal record, reconciling conflicting evidence, or grounding the conclusion in the chart in front of them. Right answer, wrong process. In some domains, that's a curiosity. In medicine, it's the entire question. A clinician doesn't just need the conclusion — they need to know the conclusion was earned. This is the gap between a model that tests well and a system a clinician can rely on. Closing it determines which healthcare AI actually reaches patients, and which stays in the pilot phase forever. The industry has to change what it measures. Accuracy alone is not a sufficient standard for clinical AI. Process fidelity, evidence grounding, and knowing when to abstain belong in the score. For anyone deploying today, value depends on: Which tasks the agent handles reliably — measured, not assumed. And how carefully it's wired into the workflow: governed data access, your own protocols enforced, evidence requirements made explicit, and a clear rule for when it defers to a clinician. Real clinical work takes more than medical knowledge. Our benchmarks have to reflect that. If you're evaluating clinical AI right now — what are you measuring beyond accuracy? Learn more at https://lnkd.in/egsXUx4zCliniCARE-Bench: Clinical AI Agents Can Be Right for the Wrong ReasonsCliniCARE-Bench: Clinical AI Agents Can Be Right for the Wrong Reasons
-
Tuan Hoang reacted on thisTuan Hoang reacted on thisI’m cracked out on Claude and thought it’d be fun to give a list of reports and tasks I’ve had it run in the last 30 days: -Business P&L -Personal P&L -Property portfolio analysis (selling 2 units bc of it) -Weekly competitor research analyzing site changes -A few deal briefs (long sales cycle email threads) -Business distribution tax strategy -Claude Marketers → substack email imports -Various Fathom sales call recaps -Graphic creation for founder-led social post -Wedding vendor coordination (yes, getting married!) Even though I write mostly about Claude helping with marketing, I’m heavily using it on the personal side too This has easily saved me 40+ hours of manual work that I likely would have put off for more important tasks
-
Tuan Hoang liked thisTuan Hoang liked thisOne of the most talented people on my CX team at Clay is an anthropologist. Her name is Gabrielle. She has a PhD in anthropology from Cornell, spent years working in the field, and then did a stint in consulting. Clay is her first job in tech. The first time I talked to Gabby, I was sick at home and on back-to-back meetings. Varun called me and my stomach dropped. A call from the co-founder on a sick day made me think something was on fire. I picked up, and it was him and Gabby on speakerphone. "Jess, I have someone I want you to meet. This is Gabby. You should interview her." That was the whole introduction (and if you know Varun, this is very on brand). Gabby inspires me because she never left her "old life" at the door. When people switch careers, they typically treat it like a blank slate. But Gabby brought everything she learned as an anthropologist into this job and made it her edge. Anthropology runs on curiosity and pattern-finding. You just keep going until you understand what you're looking at. Gabby works with customers the same way. She digs for what makes a customer's business unique and she doesn't stop until she finds their best GTM angle. She'll teach customers things about their own business they never noticed. Gabby's customers love her for it. Many of them have been promoted or won awards for work they did with Clay. Her procurement contacts have bought gifts off her wedding registry, donated to her NYC Marathon fundraiser, text her like a friend, and send her champagne when a renewal closes (see pic below) - which is crazy, because procurement's job is to usually negotiate you down. I want to hire more people like Gabby. People who bring their whole self to work and turn it into their super power. If that sounds like you, check out our open CX roles here: https://www.clay.com/jobs
-
Tuan Hoang reacted on thisTuan Hoang reacted on thisIf you are interested or know anybody who might be a good fit, DM me 🚀🚀🚀
-
Tuan Hoang reacted on thisTuan Hoang reacted on thisI'm hiring / calling for collaborators at Scale AI! If you're a researcher and interested in teaching machines intuition as a proxy for safety, Eyon Jang and I would like to hear from you. Please send me a DM if: » conduct(ed) research in AI Control/CoT monitoring, Scalable Oversight, Multi-agent Safety, Alignment Evals, Misalignment Generalization » CS/ML/Stats/Math PhD candidate (masters OK) » you can execute under fairly high-level guidance Read more about us at labs.scale.com.
Experience & Education
Languages
-
Vietnamese
Native or bilingual proficiency
-
English
Full professional proficiency
Other similar profiles
Explore more posts
Explore top content on LinkedIn
Find curated posts and insights for relevant topics all in one place.
View top content