AWS Builder Center
Full Stack Challenge: PulseMap

Full Stack Challenge: PulseMap

Vision and What It Does

Cloud Infrastructure Engineer
My challenge word was "Ripple."
The image that stuck with me: drop one stone a missed night of sleep, an extra glass of water, a ten-minute walk and it doesn't stay contained.
It spreads through everything else: your energy the next day, your mood, your focus, whether you reach for coffee or a walk.
PulseMap is a wellness tracker built around that idea. Instead of showing you a wall of disconnected numbers (blood pressure here, sleep there, mood somewhere else), it aggregates everything you log into a single Ripple Score and renders it as literal concentric rings one ring per metric, sized and colored by how it's trending. The pitch: "See your body's patterns before they become problems."
It's for anyone managing more than one health thing at once chronic conditions, cycle tracking, meds, fitness who's tired of opening five different apps to get a fragmented picture of "how am I actually doing."

Full Stack Breakdown

This is where I'll be upfront: PulseMap's frontend started from a health-tracking app I'd already been building for myself before the challenge a private, local-only tracker for blood pressure, blood sugar, fitness, hemoglobin, thyroid, and cycle tracking. When "Ripple" came up, I realized the shape of that app log an entry, see a trend, get a status was already the shape of "log daily micro-habits and watch them cascade." Here's how the project traces through every stage of the series:
  • Pitch: The elevator pitch was the Ripple metaphor itself health signals aren't isolated, they cascade so instead of five separate dashboards, one Ripple Score and a set of concentric rings tells you "how am I actually doing" at a glance.
  • Prototype: The prototype was already sitting on my machine, a personal local-only tracker I'd been building pre-challenge. That earlier work meant the core loop (log an entry → see a trend → get a status) was already validated before the 48 hours even started.
  • MVP: Scoping the MVP meant picking the smallest new slice that proved the Ripple idea: the ring visualization, a new "Daily Pulse" habit metric (sleep/hydration/movement/mood), and a rebrand — while deliberately cutting a Bedrock journaling feature to "what's next."
  • UX: UX came from a config-driven metric system one data object per metric describes its fields, chart, and status logic, and every form, chart, and history table renders generically from it, so Daily Pulse shipped with almost no new UI code.
  • Launch: Launch was the real leap for this project turning a frontend-only tool that only worked in one browser, on one device, with no backend at all, into a deployed, authenticated, cloud-synced full-stack AWS application with real accounts.

How I Built It

The frontend is React 19 + Vite + Tailwind CSS v4, using a config-driven metric system each health metric (including the new habit tracker) is just a data object describing its fields, chart, and status logic, and the UI (forms, charts, history tables) renders generically from that. Adding the whole "Daily Pulse" feature was mostly writing one config object, not new UI code.

The biggest design decision was how to sync to the cloud without a rewrite. Rather than build per-metric REST endpoints, I kept the app's existing local-first architecture (localStorage-backed, offline-friendly) and added a whole-document sync: one DynamoDB item per user holding the entire exported dataset, pulled on sign-in and pushed (debounced) after every local write. If the network or API is unreachable, the app just quietly falls back to local-only behavior nothing breaks.
The trickiest moment: my first deploy attempt included CloudFront for CDN hosting, and it failed with an account-verification error new AWS accounts need AWS Support to clear them for CloudFront resource creation. Rather than block on that, I fell back to plain S3 static website hosting, which needed no such verification and got the app live within the window. CloudFront is a clean follow-up once that clears.

AWS Services / Architecture


  • Amazon Cognito — email/password auth, JWT issuance
  • Amazon API Gateway (HTTP API) — Cognito JWT authorizer gating a  /data  endpoint
  • AWS Lambda (Node.js) — GET/PUT handlers reading and writing the user's synced blob
  • Amazon DynamoDB — one item per user, on-demand billing
  • Amazon S3 — static website hosting for the built frontend
  • Infrastructure is defined as code with AWS SAM ( template.yaml ), so the whole backend deploys with one command.

What You Learned

  • How much you can get from designing one generic, config-driven UI pattern instead of one-off screens per feature — the habit tracker was nearly free.
  • SAM's JWT authorizer on HTTP APIs is genuinely simple to wire up — far less ceremony than I expected for "real" authenticated infrastructure.
  • Not every AWS account has every service unlocked by default, and building a fallback path (S3 website hosting instead of CloudFront) mattered more under a deadline than getting the "ideal" architecture.
  • What's next: CloudFront once the account clears verification, and a Bedrock-powered journal feature that reads a free-text entry and tags it with detected symptoms automatically — cut from this pass for time, but the DynamoDB shape already has room for it.

Link to App or Repo

Any opinions in this article are those of the individual author and may not reflect the opinions of AWS.
Enjoyed reading this content? Let the author know!

Your likes, comments, shares, and saves help creators reach more builders.

Loading recommendations

Loading article