# 223 of 300 AI Certification seats are gone Canonical: https://brew.new/templates/vueschool/223-of-300-ai-certification-seats-are-gone Brand: vueschool.io Category: newsletter ![Preview of 223 of 300 AI Certification seats are gone](https://cdn.brew.new/email-preview-8fa375a44bdbae17-tracking_r57gy4xtqmhdrh05cenxg0nkp58dttdz-1789071878764.png) ## Email content Hi there, The best engineers were never the fastest typists. They were the ones who could deal with a messy problem, design the right system, and make good decisions with imperfect information. AI does the typing now. For the first time, those calls are the only thing left to measure. Four hours ago, the first 300 seats of the AI-Proficient Developer Certification opened at half their public-launch price. It's built to measure exactly those calls, and nothing else. 223 of the seats are gone, 77 are left tonight, and the launch price goes with the last of them. None of that is the reason to want a seat tonight, though. If this is the first you're hearing of it, it's the first hands-on, proctored, vendor-neutral examination of how well a developer directs AI on real software. You work inside a live codebase with an AI assistant you direct, a stakeholder whose ticket is vague on purpose, and a silent examiner reading every call you make. It comes from certificates.dev, the team behind the official Vue and Nuxt certifications. The AI-Proficient Developer Certification specimen, Certificate of Competence, AI Development Proficiency, a serial number assigned in order, the Cohort 1 marking, signed by the Examination Board Lead and the certificates.dev CEO What the seat actually pays for Most things sold as "AI training" right now are a subscription and a video library, but this one isn't. About 80% of it is hands-on, drill decks, sandbox labs, real codebases, trying the moves yourself under feedback until the judgment is actually yours, not something you watched happen to someone else. And Cohort 1's 300 seats are all there are at this price. The whole three weeks Day zero. Your prep package lands before week one, environment checked, pre-study mapped, warm-up drills in hand, so session one starts at full speed. Week 1, Sep 14, from author to director Workshop, 2.5 hours. Everyone runs an agent on the same small task, live, and the room compares who steered it where. You'll see why directing an agent is a different skill from writing code, and why your judgment matters more with AI, not less. There's a judgment speed round in the middle of it, AI diffs projected on screen, and the whole room votes ship, fix, or push back. Then comes the reveal, and how the room split. You'll also get context engineering, what you actually feed the agent decides what you get back, and exactly how the certification scores every call you make. Homework, about an hour a day. Directed builds of your own, plus confidence-scored calls on AI-written code. Live lab, 1.5 hours. The homework comes back here, solved live, your builds and calls compared across the room, and common stumbling blocks demonstrated and fixed on the spot. Then there's open Q&A, bring anything, including questions from the recording. Week 2, Sep 21, directing and reviewing Workshop, 2.5 hours. The stakeholder challenge opens it, everyone gets the same vague ticket and grills an AI stakeholder until it actually has edges. The best specs go up on screen, showing what the room found and what it missed. From there it's directing in practice, spotting an agent going off track before it burns an hour. Then it's reviewing AI code at the speed the exam demands, catching the bug patterns that compile and lie. You'll get the agent's own report, a done summary and the actual code, and find every place they don't match. There's the call almost nobody teaches too, recognizing when the code is clean and the right move is to ship it. And you'll correct the system, not the output, so every mistake an agent makes becomes a rule, a gate, or a check it can't repeat again. Homework, about an hour a day. Planted-trap code review, judging an agent's plan before any code exists, and spec-writing practice of your own. Live lab, 1.5 hours. The week's homework comes back as live review, make the call, then defend it out loud. There's also the security holes AI loves to create, injection, auth, leaked secrets, spotted in live examples. Then there's open Q&A again, bring anything, including questions from the recording. Week 3, Sep 28, debugging and exam readiness Workshop, 2.5 hours. The bug hunt opens it, a broken system on screen, minutes on the clock, everyone commits a root cause. You'll debug code you've never seen, state the problem in one sentence, rank your suspects, and find the one check that settles it, all before touching anything. You'll learn the named bug patterns worth knowing on sight, race conditions, stale caches, N+1s, timezone traps, then run the full workflow start to finish, ticket to a three-sentence defense. Week one's judgment game comes back as the championship round, at exam difficulty and exam speed, your score from day one sitting right beside today's. Homework, about an hour a day. Debugging drills across every named pattern, plus mixed exam-style practice pulling from all three weeks. Live lab, 1.5 hours. The final session is pure exam rehearsal, exam-like questions end to end, and a compressed run of the full workflow under time pressure. There's a group MCQ round too, the room answers together and the split goes up, right against wrong. Real challenges from the actual exam run in a mocked environment, so nothing about the format is a surprise. Then there's pacing the sitting, booking guidance for after the cohort ends, and a final open Q&A. The trial exam, and the sitting that counts Before you ever sit the real thing, you get the exam's own supervised environment, exactly like the sitting, and you can run it as many times as you need. The real window opens Sep 28, as week three starts, and you book your sitting on your own clock. Two sessions a week, about four hours live, plus an hour of drills a day between, with me on Discord every day of it. Every session is recorded and yours to keep. I lead the exam, and I'm the one training you I lead the exam's design, working from the same six-domain competency spine that runs through the training and the score report, and I'm also the one facilitating your live cohort in September. I spent years watching developers submit code they were completely sure was right, clean, confident, done, and the bug usually wasn't even hard to find, they just hadn't looked, because they didn't think they needed to. That gap between how sure you are and how right you actually are is why this exam makes you declare your confidence on every call, not just whether the call was correct. AI didn't create that gap. It just made it faster and a lot more common, because the confident-looking code is the AI's now, and you're still the one who has to catch it. What you actually leave with More than 600 confidence-scored judgment calls, banked across three weeks, more than most engineers get from years of on-the-job trial and error. Every session is recorded and yours to keep. And by the time you get there, the sitting is so rehearsed the exam itself is the boring part. The exam grades exactly what you train on in these three weeks, so when you pass, the six-domain score report behind it is something an employer can actually check. Every developer's resume already claims they're good with AI. This is what makes that claim provable. Deciding tonight isn't the final decision either way. The refund window runs from today until 30 days after the cohort starts, so you'll be deciding after real live sessions, not off a sales page. If you join and decide it's not for you, one email gets your money back. 114 seats left, still at half price, and the price goes with them when they're gone. Reserve my seat at half price → Alex G Technical Education Lead, BitterBrains If the AI certification isn't for you, mute these launch emails in one click. Everything else you get from us stays. -- You're getting this because you signed up at vueschool.io, part of BitterBrains – Making Developers' Lives Easier. 2093 Philadelphia Pike #9442, Claymont, DE 19703, United States · Unsubscribe | Manage preferences [Open and remix this design](https://brew.new/templates/vueschool/223-of-300-ai-certification-seats-are-gone) [Browse email designs](https://brew.new/browse/templates)