A portfolio is not a collection of everything you have built. It is four pieces of evidence, arranged so that someone with two minutes can tell what you can do. This plan gets you there in a month of evenings, and it works whether you write code, run campaigns or analyse data.
Week one: the one that proves the basics
Start with a small, complete, unglamorous project that works end to end. Complete beats ambitious every time, because a finished small thing demonstrates that you finish things, and half the applicants cannot show that.
- Data: clean a real public dataset and answer three specific questions with charts.
- Development: a small tool that does one job properly, deployed at a live URL.
- Marketing: a full campaign for a real local business, including the copy, the targeting and the measurement plan.
A tidy tutorial dataset shows you followed a tutorial. A messy real one shows you can handle what the job is actually made of.
Week two: the one that shows depth
Go one level deeper in the area you want to be hired for. This is the project you will be questioned about, so it should contain at least one decision you can defend for five minutes: why this approach, what you tried first, what you would do differently.
Week three: the one that shows range
Pick something adjacent that proves you are not a single-tool person. An analyst adds a dashboard someone else can use. A developer adds tests and a deployment pipeline. A marketer adds the landing page and the tracking, not only the ads.
Week four: the one with someone real on the other end
Do something for an actual person or organisation, even unpaid and even small. A college society, a family business, an open-source project with an open issue. This is the project that changes the conversation, because it comes with a constraint you did not choose and a person who wanted the outcome.
- Find an organisation with an obvious, small problem.
- Offer to fix that one thing, with a defined scope and end date.
- Deliver it, and ask for two lines you can quote.
The write-up is half the value
An unexplained project is close to worthless, because nobody will read your code or your campaign settings to work out what you did. Every project needs the same short page:
| Section | One or two sentences on |
|---|---|
| The problem | What was wrong, and for whom. |
| What you did | The approach, and the main decision you made. |
| The result | What changed, with a number if there is one. |
| What you would change | The honest limitation. This is the section that reads as senior. |
Add one screenshot or chart at the top. Most people decide whether to read further from the image.
Where it lives
- One page listing all four, each with a one-line summary and a link.
- Links that work without logging in, from a phone.
- The same four projects named on your resume and your LinkedIn, so the story is consistent.
Key takeaways
- Four projects: basics, depth, range, and one for a real person.
- Finished and small beats ambitious and abandoned.
- Every project needs a write-up with problem, approach, result and limitation.
- Check every link works on a phone, logged out.
Build the portfolio with people around you.
Udaan 90 is ninety days of structured, project-led learning with mentors and a cohort.
See Udaan 90