# Vibe Coding Like a Software Engineer > An online, self-paced AI course from ARFA, the TIBLOGICS AI Academy. Intermediate level, about 24 hours, in English and French, with a verifiable certificate. Web page: https://tiblogics.com/learning-box/vibe-coding-engineer ยท All tracks: https://tiblogics.com/learning-box.md ## Key facts - Level: Intermediate - Time: about 24 hours including hands-on work (7 modules, 29 lessons, 8 labs); self-paced, no deadline - Price: $487 one time for lifetime access to this track, or every track for $89 a month (cancel anytime) - Certificate: TIBLOGICS Certified AI-Assisted Software Builder, with a public verification page - Languages: English and French ## About AI can write code faster than anyone. It cannot decide what to build, notice what it broke, or keep your users' data safe. That part is engineering, and this track teaches it to people who build with AI. You will learn to write a spec before you prompt, build in small loops with version control as your undo button, read and debug code you did not write, test what the AI produced instead of trusting it, spot the security mistakes AI makes most often, and ship and maintain an app without it falling over. Security gets special emphasis, because most people who build with AI are not traditional developers. The Ship Safe module walks you through 30 doors an attacker will try before your first user signs up: keys and code history, logins and access, untrusted input, AI and agents, and what happens when it breaks. For each door you learn a two-minute check and a prompt that gets your AI assistant to audit and fix it, then you run a full security audit on a planted-flaw app and on your own. You cannot finish the track without it: the final exam covers it and your capstone needs a security section. It is hands-on from the first lesson. Every lesson has live code playgrounds and prompts you run on the page, and the labs happen in the built-in Code Studio: an editor, a live preview, an AI pair programmer whose changes you review before applying, saved versions like Git commits, and automated checks. The capstone is a small real app you build and ship yourself, reviewed by a person. Skills this track builds also appear in AI-assisted and agentic coding courses from model providers and developer platforms. This track is independent: it is not affiliated with any of them and is not official preparation for any certificate. Who it is for: Founders, analysts, designers, product people and junior developers who build software with AI tools and want to do it like a professional. No prior coding experience needed, but you should be comfortable learning by doing. ## What you will be able to do - Map the system around an app before building it: users, data, hosting and who maintains it - Write a build-ready spec with user stories and acceptance criteria an AI can build from - Build in small, reviewed steps with version control, and debug with AI methodically - Test AI-written code with edge cases and automated checks, and review its changes like a senior engineer - Find and fix the common security mistakes: exposed keys, injection, XSS and made-up packages - Run a 30-door security audit before launch, from keys and access checks to AI spending caps, prompt injection and tested backups, and fix the gaps with your AI assistant - Deploy, monitor, cost and hand over an app, and know when to bring in a professional engineer ## Curriculum ### Module 1: Vibe Coding, Engineered What vibe coding is, where it breaks, and the engineer's mindset: see the whole system around an app, pick the right kind of AI tool, and read enough HTML, CSS and JavaScript to follow what the AI writes. - What vibe coding is, and where it breaks (24 min) - Code is the smallest part of the system (25 min) - Your toolkit: chat assistants, code editors, agents and app builders (24 min) - Your first build: HTML, CSS and JavaScript in ten minutes (28 min) - Module quiz ### Module 2: Specs Before Prompts Turn an idea into something an AI can build reliably: user stories with Given/When/Then acceptance criteria, small vertical slices, a one-page spec, and a deliberately simple stack and data model. - Requirements, user stories and acceptance criteria (25 min) - Breaking work into small, testable steps (24 min) - Writing a spec an AI can build from (27 min) - Choosing a simple stack and a data model (25 min) - Module quiz ### Module 3: Building in Small Loops Build like an engineer: one change at a time, run and checked before it is saved; read and review code you did not write; use Git as your undo button; and debug by reproducing, isolating, fixing and explaining. - The build loop: one change, run it, check it, save it (24 min) - Reading code you did not write (25 min) - Git as your undo button (27 min) - Debugging with AI: reproduce, isolate, fix, explain (28 min) - Module quiz ### Module 4: Testing and Verification Prove that what the AI built actually works: manual test plans, acceptance tests and edge cases, automated tests you have checked can fail, and reviewing every AI change like a senior engineer. - "It runs" is not "it works" (24 min) - Acceptance tests and edge cases (25 min) - Automated tests with AI, verified by you (27 min) - Reviewing AI changes like a senior engineer (25 min) - Module quiz ### Module 5: Security, Data and Quality Keep secrets out of code, treat every input as untrusted, check packages before you install them, and handle logins, permissions and personal data with proven tools and the least access needed. - Secrets and API keys (24 min) - Untrusted input: injection and XSS (27 min) - Dependencies and hallucinated packages (24 min) - Logins, permissions and personal data (26 min) - Module quiz ### Module 6: Shipping and Maintaining Deploy through separate environments with a rollback plan, watch the live app with logs and monitoring, keep costs and limits under control, and document and hand over well, including knowing when to call an engineer. - Deploying: environments, configuration and going live (26 min) - Logs, errors and monitoring (25 min) - Cost, performance and limits (25 min) - Documentation, handover and knowing when to call an engineer (26 min) - Module quiz ### Module 7: Ship Safe: Security Before Your First User The most practical module in the track. Thirty doors an attacker will try on any app built with AI, grouped into five checks: before you push, auth and access, input and data, AI and agents, and when it breaks. For each door you learn why it matters, how to check it in two minutes, and a prompt that gets your AI assistant to audit and fix it. You finish with a launch audit of your own app. - Why vibe-coded apps get hacked (32 min) - Auth and access: who can open which door (34 min) - Input and data: never trust what comes in (34 min) - AI and agents: when the model holds the keys (35 min) - When it breaks: errors, logs, backups and your launch audit (32 min) - Module quiz ## How you are assessed - A quiz after each module (80% to pass, retakes allowed). - A timed final exam: 35 questions in 60 minutes, 75% to pass, 90% for distinction, 3 attempts. - A capstone project reviewed by a person against a published rubric (70% to pass). - Each certificate has a public page at https://tiblogics.com/certificates/ that anyone can open to check it.