Projects
Six projects that build on each other, each one a portfolio piece. Together they take about 122 hours. Your plan decides which ones you do.
- Project 1 AI writing assistant inside an existing app Add an AI writing assistant to an existing app with editable suggestions, accept and reject controls, and tracking of what users keep. PromptingStreaming UIEditable outputsFeedback events Level: Beginner About 10 hours
- Project 2 Smart catalog search with RAG and citations Build natural-language search over a product catalog with retrieval, cited answers and a relevance eval set that proves it beats keyword search. RetrievalCitationsGolden datasetsRelevance evals Level: Intermediate About 16 hours
- Project 3 AI onboarding agent with approvals Build an agent that sets up a new account through tools, asks for approval before risky actions, and logs every step it takes. Tool useApproval flowsGuardrailsTracing Level: Intermediate About 18 hours
- Project 4 Feedback-driven feature: mine edits, improve prompts Capture how users edit AI output, mine the edits for patterns, improve the prompt, and measure the lift with evals and product metrics. Implicit feedbackEdit analysisPrompt iterationMeasuring lift Level: Advanced About 18 hours
- Project 5 A/B-tested AI feature with a cost-per-success dashboard Ship an AI feature behind flags, A/B test two variants with analytics events, and build a dashboard of cost per successful task. Feature flagsA/B testingAnalytics eventsCost dashboards Level: Advanced About 20 hours
- Project 6 Capstone: a paid AI SaaS feature end to end Take a paid AI feature from spec to launch with UX, evals, staged rollout, pricing, a metrics review and a written case study. SpecsAI UXEvalsRolloutPricingCase studies Level: Expert About 40 hours