Capstone Index

Every deliverable from the FlyRank AI Fluency and Backend Engineering track, in one place — repos, weekly assignments, and the retrospective. Status is shown as it actually stands, not as finished-by-default.

Retrospective — written for the person I was in Week 1: what changed, what I'd build next.

Read it

Honest status note: this index links what's real as of today rather than marking everything complete. Week 8 ("The Plan to Keep Building") is drafted and ready to submit; everything else in the General AI Fluency track, including FL-09, has been submitted.

Backend Engineering Track

Each entry is a real GitHub repo with its own README and, where applicable, a live demo.

tasks-api · BE-04

Docker/PostgreSQL/Redis CRUD API — ~724x query speedup benchmarked with EXPLAIN ANALYZE on 500k rows.

Live GitHub

auth-api · BE-03

Supabase Auth + FastAPI integration.

Live GitHub

polite-scraper · BE-05

Fetch → validate → store pipeline against books.toscrape.com, with 10 offline unit tests.

Live GitHub

background-job · BE-06

FastAPI + Inngest background-job pattern with status polling, a cron heartbeat, and idempotency/concurrency/durable-restart stretch goals.

Live GitHub

llm-api · BE-07

FastAPI LLM endpoint with strict schema, timeout, retries, cost logging, a kill switch, and 8 hand-labeled eval cases.

Live GitHub

pdf-report-generator · BE-08

FastAPI/SQLite/Playwright PDF report generator — background job, cron, and email delivery, all three stretch goals done.

Live GitHub

ai-decision-flow · BE-09

Next.js + React Flow + Inngest visual AI decision-flow builder — each node is a real, durably-executed LLM call replayed live on canvas.

Live GitHub

flyrank-capstone-widget-platform

Backend Track capstone — multi-tenant embeddable widget & lead-capture platform: CORS submission API, rate limiting, spam control, geo-enrichment fallback chain. Phases 1–2 done, 24 passing tests; live Supabase + Docker/Postgres verification still pending on my own machine.

In progress GitHub

first-api · W2·A1

In-memory CRUD API — correct response shapes, 404 handling, global exception handlers.

Live GitHub

first-api-db · W3·A1

SQLite migration of first-api.

Live GitHub

General AI Fluency Track

Most weeks fed directly into this site. Where a week doesn't have a public artifact of its own, it's a FlyRank portal submission.

FL-01 — Workflow audit

Audit of real day-to-day AI tool use.

Done

FL-02 — Prompt engineering

Cross-model prompt comparison.

Done

W1 — Draw the Path

Portfolio sitemap (Home, Work, About, Contact) + this Claude Project set up.

Done Home

W2 — Frame It as Cases

Voice card + three case studies — became the Work page.

Done Work

W2 — Prompt Ladder

Prompt-iteration exercise.

Done

Make It Do Something

One working dynamic feature on a free tier — the Web3Forms-backed contact form, real test submission confirmed.

Done Contact

Open It on Your Phone

Mobile/tablet/desktop QA pass — real fixes: compressed background image, iOS zoom-on-focus bug, stale copy.

Done

Survive the Crit

Real reviewer feedback against the site's proof statement — positioning landed; more case studies flagged as the nice-to-have.

Done

Break Your Own Site

Edge-case + hardening pass — found and fixed a genuinely broken Contact link, added OG/Twitter meta, robots.txt, sitemap.xml.

Done

Plant Your Flag

Custom domain (safva.dev), free analytics, and FlyRank graduate badge.

In progress

The Plan to Keep Building

How-to-add-the-next-case note, ai-decision-flow named as the next case study, and a real reminder set for Oct 13.

Drafted

FL-09 — Documentation and demo

SOLACE's README + a demo video, posted to the showcase thread.

Submitted Demo

FL-10 — Final package & capstone

This index, the retrospective, the hours log, and the live site.

In progress Retrospective

Portfolio site

Home

LiveOpen

Work

LiveOpen

Contact

LiveOpen

About

Not started

← Back to home