A Next.js web app that helps you master concepts for AI Product Managers through structured lessons, spaced repetition flashcards, AI-powered quizzes, and visual reference guides.
- 22-topic curriculum across 4 pillars: Technical Foundations, AI Product Craft, Strategy & Business, and Safety, Ethics, and Governance
- 21 reference guides — shorter, visual explainers (diagrams, tables, charts) for topics that don't fit a single quiz question
- Learn Mode — Read through structured lessons with key takeaways
- Review Mode — Spaced repetition flashcards using the SM-2 algorithm (159 cards)
- Quiz Mode — LLM-generated interview-style questions with AI-evaluated free-text answers
- Bookmarks — Save specific sections of a lesson for later
- Progress tracking — Track completed topics, quiz scores, and overall mastery
- No account, no database — the app works entirely as a guest; progress is saved in your browser's local storage. There's no sign-in flow and no database in the stack at all
- Dark mode — Toggle between light and dark themes
# Install dependencies
npm install
# Copy the env example
cp .env.example .env.localEdit .env.local and set:
AI_GATEWAY_API_KEY— see AI Gateway Setup below, needed for quiz generation and answer evaluation.AUTH_SECRET— optional. Sign-in is disabled in the UI, but leaving this blank produces a harmlessMissingSecretlog on the server (any random string, e.g.openssl rand -base64 32, silences it).
AUTH_URL, AUTH_GITHUB_ID, and AUTH_GITHUB_SECRET can be left blank — they're unused while sign-in is disabled.
npm run devOpen http://localhost:3000 to start learning. There's no database to set up — the app runs fully in guest mode, with progress saved to localStorage.
Quiz generation and answer evaluation use the Vercel AI Gateway for LLM access. You need an API key:
- Go to your Vercel Dashboard
- Navigate to AI Gateway → API Keys
- Create a key and add it to
.env.localasAI_GATEWAY_API_KEY
The app uses openai/gpt-4o-mini by default. You can change the model in app/api/quiz/route.ts, app/api/evaluate/route.ts, app/api/post-lesson-mcq/route.ts, and app/api/quiz-example-answers/route.ts. These routes are rate-limited per IP to guard against abuse from anonymous guest traffic (see lib/rate-limit.ts).
| Pillar | Topics |
|---|---|
| Technical Foundations | ML Fundamentals, Deep Learning, NLP & LLMs, Foundation Models & RAG, Computer Vision, Recommender Systems, Agentic AI |
| AI Product Craft | AI Product Design, Human-in-the-Loop, AI Evaluation & Metrics, Prompt Engineering, AI UX Patterns |
| Strategy & Business | AI Product Strategy, Business Cases & ROI, Data Strategy, MLOps |
| Safety, Ethics, and Governance | AI Safety & Alignment, Ethics & Bias, Regulation & Compliance, Explainability & Interpretability, AI IP & Copyright, Responsible AI |
Shorter visual references that don't warrant a full lesson + quiz — model selection, prompt & policy design, RAG vs fine-tuning vs prompt-only, retrieval architecture, evaluation strategy, pricing & unit economics, caching & request shaping, embeddings vs keyword search, latency & streaming UX, agent boundaries & HITL, tool & integration design, hallucination & grounding, AI security basics, AI governance & data lifecycle, observability, experimentation & rollout, multi-tenant AI SaaS, vendor & routing strategy, working with ML teams, synthetic data & augmentation, and model monitoring & drift.
- Next.js 16 with App Router and TypeScript
- Tailwind CSS v4 + shadcn/ui
- Vercel AI SDK with AI Gateway
- SM-2 spaced repetition algorithm
In Review mode:
- Space / Enter — Flip card
- 1 — Again (forgot)
- 2 — Hard
- 3 — Good
- 4 — Easy