Shipping, and What Comes Next

You've built it. Now make it count.

The shipping checklist

Run through this before declaring the capstone done. Five sections, ~20 items.

Repository

• GitHub repo is public.
• README has: 1-paragraph elevator pitch, architecture diagram (mermaid renders on GitHub), data sources, tech stack, how-to-run, license.
• LICENSE file present (MIT or Apache for the code; CC-BY-4.0 for the data if Project D).
• .env.example documents required env variables; no secrets in git history.
• At least one GitHub Action visible in the repo (CI tests or scheduled refresh).
• Last commit message is meaningful, not "wip" or "fix".

Deployment

• The deployed URL is publicly reachable (no auth wall for the demo views).
• HTTPS works.
• Loads in under 3 seconds on a fresh browser.
• At least 14 days of real captured data visible.
• No "broken" / "TODO" buttons or links on visible pages.

Documentation

• Blog post is written and published (your blog, Medium, or scrapingcentral.com/blogs).
• Blog post explains: what + why + how + what broke + what it cost. ~1500–2500 words.
• Cost breakdown is honest: actual monthly figure, including domain + VPS + SERP API + proxies.
• At least three "things that broke and how I fixed them" sections.

Polish

• No emojis in formal text. Use them sparingly in casual sections (or not at all).
• All units labelled (USD vs INR; sqm vs sqft; UTC vs local).
• All charts have axis labels.
• Mobile rendering works (a single break in your dashboard is enough to disqualify you for a UI-conscious team).
• No console errors on the dashboard.

Reach (the bit most students skip)

• Post on LinkedIn: screenshot, one paragraph, GitHub link, deployed link.
• Post on X / Bluesky / your home community with the same.
• Submit to relevant awesome-lists, subreddits, niche newsletters.
• Email three people who'd plausibly care.
• If applying for jobs: link the capstone in your CV and cover letter.

The first 30 days after shipping

Do these in order:

Days 1–3: Announce

• Post on social. Use the screenshot you took at the moment of shipping; that's the photo.
• DM five people who'd plausibly care. Personal beats broadcast.
• Submit to awesome lists.

Days 4–10: Apply

• Apply for five scraping-related jobs. Link the capstone in the application.
• Send three cold pitches to companies you'd like to scrape for. Format: "Saw you have X data on your site; I scrape Y for Z (link to capstone); want to chat?"
• Update your LinkedIn headline + project section.

Days 11–20: Operate

• The deployed instance will break. Something will go wrong.
• Fix the first failure. Write up the fix on a blog/Twitter thread.
• Add a status badge / health-check page so users can see the project is alive.
• Note the second failure for a future blog post; fix it.

Days 21–30: Reflect

• Open a follow-up issue in the repo: "lessons from operating 30 days."
• If anyone has used the project, ask them what's missing.
• Decide whether to: (a) double down (add features, charge money), (b) move to a second project, (c) maintain quietly.

Don't start a second project before day 30. The first month is when the first project produces its compounding returns; jumping to a second one too early loses the leverage.

What comes next: three paths

After the capstone, scraping engineers diverge into three rough paths.

Path 1: Employment

You want a salaried scraping engineering role. The capstone is your strongest signal:

• Apply via LinkedIn, AngelList, niche scraping companies (Bright Data, Zyte, Apify, the SERP API providers all hire).
• Apply via your network. The "I built X" message converts better than "I'm looking for a job."
• Expect interview takehomes that look like a smaller version of your capstone. You'll crush them.
• Salary range (2026, rough): $50k–$120k entry, $90k–$180k mid, $150k+ senior, varies wildly by location. Adjust for your region.

Path 2: Freelance / consulting

You want to charge per project.

• Set up a tiny site (yourname.com): three pages, home, projects (link the capstone), contact. WordPress / Astro / hand-rolled HTML, doesn't matter.
• Post in scraping subreddits + Discord servers when projects come up.
• Charge $50–100/hr to start; raise as portfolio grows.
• A single capstone-shaped project can quote at $3,000–10,000 fixed, depending on scope.
• Cover yourself: contracts with explicit data-usage clauses, kill switches.

Path 3: Productised / SaaS

You want recurring revenue from one of your scraping skills.

• Project E (rank tracker) is the obvious productise-able capstone.
• Project A (price intelligence) productises into a Shopify app or a niche tracker.
• Project D (open data) productises into a freemium API or a "premium aggregated CSV" tier.
• Start at $19–49/mo per user; aim for ten paying users in six months.
• Stripe integration takes a weekend; lean on it.

The lines blur. Many scraping engineers do all three at different ratios over time.

A few honest things

• Your first capstone won't be your best work. That's the point. The second is better. The fifth is what employers see.
• Most attempts stall in week 3. The fix is to ship at 70% complete instead of waiting for 100%. A live janky thing beats a perfect imagined thing.
• No one cares as much as you do. Your friends won't read the blog post. Hiring managers will skim it for 90 seconds. Make those 90 seconds count.
• The curriculum gave you the toolkit, not the intuition. Intuition comes from operating the deployed instance for 30 days while it breaks in interesting ways.

What you've actually done

If you got to here:

• You went from "what is HTTP?" to operating a multi-source production scraper.
• You shipped working code in two languages (Python + PHP).
• You did the boring work of normalising data, monitoring, deploying.
• You wrote about it publicly.

That's a job's worth of skill. Most CS degrees produce graduates who can't do this. You can. Go act like it.

Hands-on lab

Open your capstone repo. Check off the shipping checklist above. Anything unchecked: fix it today, or write it into a follow-up issue with a deadline. The deadline matters. The issue alone doesn't.

You're done.