The Rules of This 30-Day Sprint
Thirty days, thirty images, one portfolio site, four weekly posts. That is the whole contract. You will not master every tool, you will not nail every render, and you will absolutely make ugly things in week one. None of that matters. What matters is that on day 30 you have a public URL with a body of work, a clear visual point of view, and the muscle memory to keep going.
Three non-negotiables before you start:
- Pick one primary tool and stick with it for the full month. Midjourney, Flux, or SDXL — choose based on chapter 3, then commit. Switching mid-sprint resets your prompt intuition.
- Pick one theme. "Cyberpunk food photography," "isometric tiny worlds," "1970s editorial portraits of fictional musicians" — narrow enough that ten images in, your style is recognizable. Generic "cool AI art" is not a theme.
- Ship daily. A bad image shipped beats a great image stuck in your drafts. The portfolio is a deadline, not a museum.
If you have not finished /courses/ai-image-generation-beginners, do it this weekend before day 1. The sprint assumes you can prompt, iterate, and inpaint without looking things up.
Week 1: Foundation (Days 1-7)
The goal this week is volume and self-honesty. You are calibrating your eye and your prompt vocabulary.
- Day 1: Generate 10 images on your theme. Pick your favorite. Write down why in one sentence.
- Day 2: Recreate yesterday's favorite from scratch without looking at the prompt. Compare. What did you forget?
- Day 3: Push composition. Three images using rules from chapter 6 — rule of thirds, leading lines, low angle.
- Day 4: Push style. Three images referencing a specific photographer, illustrator, or film. Cite them in your notes.
- Day 5: Push mood. Same subject, three lighting setups: golden hour, harsh noon, neon night.
- Day 6: Pick your best image of the week. Inpaint to fix the one thing that bugs you.
- Day 7: Write the week-1 post. Show three images, what you tried, what failed. Publish it even if the site is not built yet — a public Are.na or X thread counts.
Checkpoint: do you have seven images you would not be embarrassed to show? If three or fewer, your theme is too broad. Tighten it before week 2.
Week 2: Build the Site (Days 8-14)
Half your week is portfolio construction. Use whatever ships fastest — Cargo, Framer, a Next.js template, a single static HTML page on GitHub Pages. Custom is not the goal; visible is.
Site requirements:
- Homepage grid of 12-20 images
- One project page per theme
- About page: 2-3 sentences + contact
- Each image: alt text + one-line caption
- AI disclosure note (see chapter 11)
- Days 8-9: Build and deploy the empty shell. Custom domain optional but recommended.
- Day 10: Curate. Go through week 1 images, pick the five strongest. Re-export at full resolution.
- Days 11-13: Generate one hero image per day specifically sized and composed for the site — wide landscape for the header, square for the grid, vertical for the case study. Treat the site itself as a creative brief.
- Day 14: Publish. Post the link. Write week-2 post: behind the build, three favorite images, one failure.
Checkpoint: a stranger lands on your homepage. Within five seconds, can they describe your theme in one sentence? If not, your grid is too varied. Cut, do not add.
Week 3: Depth (Days 15-21)
Volume gave you reps. Now you go deeper on three pieces.
Pick three concepts you want to develop into proper projects — a series, a poster set, a fake editorial spread. Each gets two days.
- Days 15-16: Project A. Five iterations, inpaint refinement, final export at print resolution. Document the prompt evolution.
- Days 17-18: Project B. Same process. Force yourself into a different style than A.
- Days 19-20: Project C. Apply character or product consistency from chapter 9. This is the one recruiters and clients care about.
- Day 21: Add all three projects as case study pages. Write week-3 post: one project breakdown, prompt included.
Rubric for each finished project, score yourself honestly out of 5:
- Concept clarity: does the image say one thing well?
- Composition: would it survive being cropped to a thumbnail?
- Technical: any artifacts, mangled hands, weird text?
- Style fit: does it belong with the rest of your portfolio?
- Shareability: would you post this without a disclaimer?
Anything under 3 on any axis goes back into iteration, not the portfolio.
Week 4: Ship and Position (Days 22-30)
The work exists. Now make it findable.
- Day 22: Real projects. Make a thumbnail, slide deck cover, or book cover (see chapter 10) as if a client briefed you. Add to portfolio.
- Day 23: Repeat for a different format.
- Day 24: Write one tutorial post: pick your best image, walk through the prompt, the iterations, the edits. This is what gets indexed and shared.
- Day 25: Audit. Remove three weakest images from the site. Quality beats quantity at the portfolio stage.
- Day 26: SEO and metadata. Title tags, OG images, descriptions. Submit to one directory or community.
- Day 27: Outreach. Email five people whose work inspired your theme. Show them yours. No ask, just connection.
- Day 28: One ambitious final image. The cover of your portfolio. Spend the full day.
- Day 29: Final post: 30 days, what you learned, three favorite images, what you are doing next.
- Day 30: Rest. Look at day 1 next to day 30. Notice the gap.
What You Have on Day 31
A live portfolio site with at least 20 curated images, three case studies, four public posts, and a documented prompt library. More importantly: you know your taste, your tools, and how long things take. That last one is what separates hobbyists from people who get hired.
Keep the cadence. One image a day, one post a week, one project a month. The portfolio is not the finish line — it is the receipt that proves you can ship.