§07ship, don’t scroll

Three projects. One certificate each.

Every project has a spec, a starter, a rubric, and a Claude reviewer that grades your submission against it. Ship one — get a portfolio-grade certificate you can share on LinkedIn.

Beginner+1–2 hours

Scrape every Python question from Hacker News

Your own weekly Python-news archive.

Fetch HN's Python tag feed, parse the front-page entries, dedupe by URL, and save the result to a JSON file. Run it weekly to build an archive of Python conversations you can search.

Skills

requests · BeautifulSoup · file I/O · dict / list handling

Open spec→

Intermediate2–3 hours

Analyze a year of your spending

Turn your CSV bank statement into a leak-finder.

Load a CSV bank statement, group by month + category with pandas, and plot the trends with matplotlib. Deliver a `spending.png` chart and a printed summary showing your top three leak categories.

Skills

pandas · matplotlib · CSV parsing · dataclasses

Open spec→

Advanced3–4 hours

A Telegram bot that summarises any URL

Async, decorator-driven, deployable in an hour.

Build a Telegram bot that listens for URLs in your DM, fetches the page, sends a 3-line summary back via Claude. Async, uses decorators for message routing, and ships on Railway's free tier.

Skills

async / await · decorators · python-telegram-bot · Anthropic SDK · environment variables · type hints

Open spec→

howthe loop

step 01
Read the spec
Goal, rubric, starter code — all on the project page. No filler.
step 02
Ship it
Push to a public GitHub gist or repo. Deploy if the project needs a runtime.
step 03
Get reviewed
Paste your URL — Claude reads your code against the rubric and returns Markdown feedback.
step 04
Earn the cert
Once the rubric is green, your public per-project certificate goes live. Share it anywhere.

context · portfolio

Why these three projects, and not another to-do app

Every Python tutorial ends with a to-do CLI. No recruiter has ever hired for that. We picked three projects that mirror what a junior Python role at a product company actually looks like: an HN scraper (teaches requests, BeautifulSoup, rate limits, pagination — the bread and butter of any data role), a spending analyzer (pandas groupby, date parsing, matplotlib — the exact stack in an analyst’s day one), and a Telegram bot (async, webhooks, deployment — how backend interns ship). Each project runs through an AI reviewer trained on a fixed rubric: correctness, idiomatic Python, error handling, and shippability. “Shippable” means it survives someone else running it on their machine — the single skill that separates a placement-ready portfolio from another tutorial-follower’s GitHub graveyard.

loading…