DEV Community

Cover image for I Built a Free AI Education Platform in 5 Languages — And Open-Sourced Everything
Ramesh Reddy Adutla
Ramesh Reddy Adutla

Posted on • Originally published at aieducademy.org

I Built a Free AI Education Platform in 5 Languages — And Open-Sourced Everything

Last year, my mother asked me what I do for a living. Not the job title — she's heard "Technical Lead" plenty of times. She wanted to understand what AI actually is. Why everyone keeps talking about it. Whether it'll take her grandson's future job.#

She speaks Telugu. I searched for a structured, beginner-friendly AI course in Telugu.

I found exactly zero.

Not behind a paywall. Not for free. Not anywhere. The same was true for my colleague's parents who speak Hindi, and a friend in Brussels who works in Dutch. If you don't speak English fluently, the door to understanding AI is basically shut.

That didn't sit right with me. So I built the door myself — and left it wide open.

AI Educademy is a free, open-source platform that teaches AI from absolute beginner to production mastery — in 5 languages, with 10 structured programs, 100+ interactive lessons, and a playground where you can experience AI before writing a single line of code.

🔗 Try it now: aieducademy.org
Star it on GitHub: github.com/ai-educademy/ai-platform


🚨 The Problem: AI Education is Broken

Let's be honest about the state of AI education in 2025:

It's fragmented. You cobble together a YouTube playlist, a Medium article, half a Coursera course, and a random Kaggle notebook. There's no coherent path from "What is AI?" to "Deploy an ML pipeline in production."

It's English-only. Over 75% of the world doesn't speak English as a first language. A curious teenager in Hyderabad, a career-changer in Lyon, a retiree in Eindhoven — they're all locked out of the best learning resources.

It's expensive. The best structured programmes cost $50–$500. In many countries, that's weeks of wages.

It assumes you're already technical. Most courses open with "Let's import TensorFlow" before explaining what a neural network actually is. If you're not already a programmer, good luck.

AI will reshape every industry and every job. The people building it — and the billions of people affected by it — shouldn't need a credit card and fluent English to understand it.


🌱 The Vision: AI Education as a Right, Not a Privilege

AI Educademy is built on three convictions:

  1. Free forever. No paywalls, no premium tiers, no "upgrade to unlock." MIT licensed. The code, the content, everything.
  2. Your language first. Every lesson available in English 🇬🇧, French 🇫🇷, Dutch 🇳🇱, Hindi 🇮🇳, and Telugu 🇮🇳 — with more languages coming from the community.
  3. Zero to hero, with no gaps. A structured curriculum that takes you from "I've never heard of machine learning" to "I can design and deploy AI systems in production."

🗺️ What I Built: The Full Tour

10 Programs, 2 Tracks, 1 Journey

AI Educademy uses a nature growth metaphor. You start as a seed and grow into a forest:

🌱 AI Seeds        → "What is AI?" (no code, no maths)
🌿 AI Sprouts      → Foundations (data, algorithms, neural nets)
🌳 AI Branches     → Specialisations (ML, CV, NLP, GenAI)
🏕️ AI Canopy       → Production AI (MLOps, RAG, governance)
🌲 AI Forest       → Mastery (research, leadership, frontier AI)
Enter fullscreen mode Exit fullscreen mode

There's also a parallel craft-engineering track for people who learn by building:

✏️ AI Sketch       → Design thinking for AI products
🪨 AI Chisel       → Engineering fundamentals
⚒️ AI Craft        → Building real AI systems
💎 AI Polish       → Production-grade refinement
🏆 AI Masterpiece  → Capstone portfolio projects
Enter fullscreen mode Exit fullscreen mode

10 programs. 2 tracks. 5 languages. 100+ MDX lessons. All free. All open source.

🌍 Full Multilingual Support

This isn't a "we translated the navbar" situation. Every lesson, every programme description, every UI string is available in all 5 languages. The next-intl integration detects your browser language and routes you automatically — or you can switch with one click.

If a lesson hasn't been translated into your language yet, the platform gracefully falls back to English. Contributors can translate one lesson at a time without breaking anything.

🎮 Interactive Playground & Lab

Learning AI by reading about it is like learning to swim by reading about water. The Playground brings AI to life:

  • Mini-games — interactive challenges that teach AI concepts through play, not lectures
  • System Design Canvas — a drag-and-drop canvas where you architect AI systems visually, connecting components like data pipelines, model training, and inference endpoints
  • Drawing Recogniser — sketch something and watch a pre-trained model guess what you drew in real time

No setup. No installation. Just open the browser and start experimenting.

📱 PWA with Offline Support

AI Educademy is a full Progressive Web App powered by Serwist (a modern Workbox wrapper for Next.js). Install it on your phone or tablet, and access cached lessons even without internet. For learners in areas with unreliable connectivity — which is much of the world — this isn't a feature. It's a necessity.

🔊 Text-to-Speech for Accessibility

Every lesson can be read aloud with built-in text-to-speech. Whether you're visually impaired, a commuter listening on the go, or someone who simply learns better by hearing — the content meets you where you are. Accessibility isn't an afterthought. It's a Lighthouse score of 100.

🔥 Streak Tracking & Progress Dashboard

Track your learning across all programs: completion percentages, daily streaks, and a visual map of your entire journey. The streak system adds just enough gamification to keep you coming back without turning education into a dopamine trap.

🤖 GitHub Copilot Extension

I built a GitHub Copilot Extension for AI Educademy so you can ask questions about AI concepts, get lesson recommendations, and explore the curriculum — all from within your IDE or GitHub. Learning AI while you're coding? That's the dream.


🏗️ Tech Stack Deep Dive

Here's the architecture for the engineers in the room.

The Foundation

Layer Technology
Framework Next.js 16 (App Router) + React 19
Language TypeScript (strict mode)
Content MDX via next-mdx-remote + gray-matter
Styling Tailwind CSS 4
Animations Framer Motion 12
i18n next-intl (routing, messages, content)
PWA Serwist
Deployment Vercel (edge-optimised)
CI/CD GitHub Actions → Vercel auto-deploy
License MIT

App Router + Locale-Aware Routing

Every route lives under a [locale] dynamic segment:

src/app/[locale]/
├── page.tsx                              # Homepage
├── programs/[programSlug]/
│   └── lessons/[slug]/page.tsx           # Lesson viewer
├── dashboard/page.tsx                    # Progress tracking
├── playground/page.tsx                   # Interactive lab
└── about/page.tsx
Enter fullscreen mode Exit fullscreen mode

The next-intl middleware detects the browser language, matches against supported locales (en, fr, nl, hi, te), and routes automatically. Users can override via a LanguageSwitcher component at any time.

Git Submodules: Content as Microservices

This is the architectural decision I'm proudest of. Each of the 10 programs is its own GitHub repo, pulled into the platform via git submodules:

ai-platform/content/programs/
├── ai-seeds/          ← github.com/ai-educademy/ai-seeds
├── ai-sprouts/        ← github.com/ai-educademy/ai-sprouts
├── ai-branches/       ← github.com/ai-educademy/ai-branches
├── ai-sketch/         ← github.com/ai-educademy/ai-sketch
└── ...                   (10 repos total)
Enter fullscreen mode Exit fullscreen mode

Why? Because content and code have fundamentally different lifecycles. A translator fixing a typo in a French lesson shouldn't need to understand the React codebase. Each content repo has its own GitHub Action that fires a repository_dispatch to the platform repo, triggering a Vercel rebuild. Content changes go live in ~60 seconds.

Want to add a new program? Create a repo, add a submodule, add one entry to programs.json. Routing, i18n, and rendering all work automatically.

Lighthouse Scores

I take performance seriously. Current Lighthouse scores:

Metric Score
⚡ Performance 90
♿ Accessibility 100
✅ Best Practices 100
🔍 SEO 92

Getting accessibility to 100 with 5 languages, text-to-speech, dark mode, and complex interactive components — that took work. But it matters.


🧩 Challenges I Didn't Expect (And How I Solved Them)

Challenge 1: i18n for Content, Not Just UI

Everyone talks about i18n for button labels and error messages. That part is easy — next-intl handles it beautifully. The hard part is internationalising 100+ long-form educational lessons across 5 languages.

Each lesson is an MDX file, and translations live in parallel directory structures (lessons/en/, lessons/fr/, etc.). I built a content fallback chain that tries the user's locale first, then degrades to English:

export function getLesson(program: string, slug: string, locale: string) {
  const localePath = `content/programs/${program}/lessons/${locale}/${slug}.mdx`;
  const enPath = `content/programs/${program}/lessons/en/${slug}.mdx`;

  const filePath = fs.existsSync(localePath) ? localePath : enPath;
  const { data, content } = matter(fs.readFileSync(filePath, "utf-8"));
  return { meta: data, content };
}
Enter fullscreen mode Exit fullscreen mode

Simple in concept, but the implications are powerful: contributors can translate one lesson at a time. The platform never breaks. You always get content — just maybe not in your preferred language yet.

Challenge 2: Making Submodules Not Terrible

Git submodules have a reputation. Partly deserved. The git clone --recurse-submodules dance trips up every new contributor.

My solution: a comprehensive CONTRIBUTING.md with copy-pasteable commands, a setup script that handles submodule initialisation, and CI that validates submodule state on every PR. I also configured Vercel's build command to always run git submodule update --init --recursive before the Next.js build. Zero manual intervention in production.

Challenge 3: PWA Cache Strategy for Multilingual Content

Caching is hard. Caching when each page exists in 5 language variants is really hard. A naive cache-first strategy would serve French content to English users if the French version was cached first.

The fix was a cache key strategy that includes the locale in every cache entry, combined with a stale-while-revalidate approach for lesson content. Serwist made the service worker configuration manageable, but getting the cache invalidation right across languages took several iterations.


🔮 What's Next

AI Educademy is live and growing, but the roadmap is ambitious:

  • More languages — Spanish, Portuguese, Arabic, and Mandarin are the most-requested. Every new language contributor multiplies the platform's reach.
  • More content — The advanced tracks (AI Canopy, AI Forest, AI Masterpiece) are being actively written. The beginner tracks are already comprehensive.
  • Community contributions — I want this to become a community-driven project where domain experts contribute lessons in their speciality and translators make them accessible worldwide.
  • Assessment engine — Interactive quizzes and projects that give learners real feedback on their understanding.

🤝 How You Can Get Involved

If you've read this far, you probably care about making education accessible. Here's how to help:

🌍 Translate a lesson

Every lesson translated is another person who can learn AI in their mother tongue. The format is just MDX (Markdown with components) — you don't need to be a developer. Pick a lesson, translate it, open a PR.

🛠️ Contribute code

The stack is Next.js 16, React 19, TypeScript, and Tailwind CSS 4. The design system has its own Storybook. If you can build React components, you can improve AI Educademy.

📢 Share it

Know someone learning AI? Someone who teaches? Someone who speaks French, Dutch, Hindi, or Telugu? Send them aieducademy.org. A single share can change someone's learning trajectory.

⭐ Star the repo

Stars drive GitHub discoverability. More stars → more contributors → more languages → more learners. It takes 2 seconds and costs nothing.

⭐ Star AI Educademy on GitHub


🔗 Links


AI is the most transformative technology of our lifetime. Understanding it shouldn't depend on what language you speak or how much money you have.

If this resonates with you — try it, share it, star it, or contribute. Every bit helps.

If you found this useful, consider giving it a ⭐ on GitHub. It means more than you think.


Built solo with ❤️ by Ramesh Reddy Adutla — Technical Lead by day, open-source education advocate by night.

Top comments (0)