
Weekend Showcase Challenge: StudySprint AI
Study Companion build with Lambda and Amplify
Vision and what it does
Studying a new topic often starts the same way: you open a browser tab, type in a subject, and get buried in links, videos, and half-finished notes before you've even figured out where to begin. StudySprint AI exists to skip that friction. You tell it what you want to learn, your current skill level, and how much time you have — and it hands you back a structured study session: a step-by-step study plan broken into timed stages, a set of concise key-concept notes, and a 5-question knowledge check to test what you've absorbed.
After the quiz, StudySprint scores you and tells you exactly what to revisit before moving on. No login, no setup, no clutter — just a topic in, a focused study session out.
How I built it
I set out with one constraint: keep it as small and dependency-light as possible, and get a fully working, deployed version before adding anything extra. That meant building the entire frontend first against mock data — a React + Vite app with the full user flow (topic input → generated plan → quiz → score → recommendation) — before writing a single line of backend code. Once the UI was solid, I moved to the backend, replacing mock data with a real AWS Lambda function exposed through a Lambda Function URL, which let me skip API Gateway entirely and keep the architecture lean.
A key decision early on was choosing Lambda Function URLs over API Gateway. For a project this size, an API Gateway adds cost and configuration overhead I simply didn't need — a Function URL gives me a public HTTPS endpoint straight from Lambda, with CORS configured natively.
The trickiest part of the build wasn't the AI logic — it was making sure the browser and Lambda could actually talk to each other cleanly, and being deliberate about which AWS services to lean on so the whole stack stayed inside the Free Tier. That constraint shaped almost every decision: no database, no authentication, no persistent storage. Each study session is generated fresh, on demand.
AWS services used and architecture
React (AWS Amplify) │ HTTPS POST ▼AWS Lambda (Function URL) │ ▼Structured JSON response │ ▼React dashboard: plan, notes, quiz- AWS Lambda — the serverless backend that validates input and generates the study session content
- Lambda Function URLs — a direct HTTPS endpoint into Lambda, skipping the need for API Gateway
- AWS Amplify — hosts and deploys the React frontend directly from GitHub, rebuilding automatically on every push
The frontend and backend are fully decoupled by design — the React app only knows it's talking to a JSON API. That means the generation logic behind that API can evolve independently, without ever touching the UI again.
What I learned this summer
This project reinforced something I'd tell any builder starting out on AWS: build in layers, and prove each one before adding the next. Mocking data before touching AWS meant I never once paid for iteration on the UI. Choosing Function URLs over API Gateway meant one less service to configure and reason about. And staying disciplined about what not to add — no database, no auth, no orchestration — kept the whole project inside the Free Tier while still telling a complete serverless story.
The biggest lesson, though, was about resilience: not every AWS service will behave exactly as expected on a tight deadline, and a good architecture is one where you can route around a blocker without rebuilding everything else.
Try it
Shoutout
A shoutout to Lewis Sawe, whose always-on museum-curating agent from the Creative App challenge was one of the most imaginative builds I saw this summer — proof that a small, well-scoped idea can still feel genuinely alive.
Enjoyed reading this content? Let the author know!
Your likes, comments, shares, and saves help creators reach more builders.
Loading recommendations
Loading article