All video posts

Portfolio · Junior developers · Career

How to Build a Capstone Project That Gets Junior Developers Hired

Through years of hiring developers and helping people switch into tech, I noticed one pattern. The people who stand out are not churning through tutorials or copying apps. They have built one real thing that adds value. Every junior developer needs a capstone project, and here is how to build a good one.

Deric YeeDeric Yee 16 November 2026 10 min read 25 min video

Prefer to read? The full article is below. Watch on YouTube.

The short version

  1. 1A pile of tutorial clones makes you look like every other applicant. One deep, original project proves you can solve problems.
  2. 2Tutorial follow-alongs, pixel-perfect clones, generic course projects, tiny freelance sites and token pull requests do not move the needle.
  3. 3Find your idea in your own life: problems from past jobs, hobbies, manual processes you could automate, or tools developers would use.
  4. 4Pick one stack and go deep, plan an MVP, use Git from day one, write a proper README and deploy it so people can click a link.
  5. 5Ship early, gather feedback and keep iterating. The difference between a dabbler and a developer is finishing.

Stop being a code monkey

A lot of junior developers get stuck in what I call code monkey mode: someone who writes code without thinking independently, follows instructions well and follows tutorials blindly. When I started landing bigger software projects and needed to hire, I was shocked at how many applicants had superficial skills and identical portfolios. A to-do app here, a weather app there, all from the same tutorials, with no original problem solving or personality on show.

Watch a video, follow along, build a clone, add it to your CV, repeat. You end up with a dozen tiny projects that all look the same, and you blend in with thousands of other newcomers. From a hiring manager's view, nothing says "I can add value to your team".

To be fair, those small projects do serve a purpose. They reinforce syntax, frameworks and how an API call works. But at some point you have to graduate from copy-paste coding and take on a project that forces you to think for yourself. That is what moves you from code monkey to problem solver, and it is exactly what a capstone project is for.

“No one hires a musician because they can play scales. They hire someone who can perform a full song.”

Deric Yee, in the video

What a capstone project is, and why it matters

A capstone is the big final piece that showcases everything you have learned: a comprehensive, end-to-end project you genuinely care about, which solves a real problem or provides real value to users. Front end, back end, deployment, the works. It is not another contrived exercise. It is something you might use yourself or offer to others.

Go back to first principles. Why does a company hire a developer? To solve problems and create value. Languages, frameworks and algorithms are tools towards that goal. If your CV only shows classroom exercises and clones, an employer cannot tell whether you can only follow instructions or can tackle an unknown problem. A project where you identified a problem, designed a solution, built it and got real feedback is direct evidence.

It also changes your interviews. Instead of generic JavaScript questions, you get asked: "I see you built a personal finance app for freelancers. What was the hardest part?" Now you are discussing trade-offs, obstacles, user feedback and scaling. You are showing how you think, not just what you know. I have been on both sides of the hiring table, and one deep dive into a real project impresses far more than a laundry list of shallow apps. More on this in what hiring managers look for in junior developers.

Portfolio pitfalls to stop wasting time on

Before we talk about building a great capstone, make sure you are not spending your time on the wrong things. These are the patterns I see most often.

01

Tutorial follow-alongs and clones

Fine for learning, but an employer cannot tell what you did versus what the instructor spoon-fed you. Pixel-perfect Instagram or Spotify clones sharpen CSS skills, but they are not your idea and do not show you can design features around user needs.

02

Course projects thousands have built

If hundreds of students built the same generic e-commerce site or library system from the same brief, hiring managers have seen it before. It will not set you apart.

03

One-week freelance gigs

A simple site for a local shop shows hustle and is worth mentioning. But a single page anyone could make with a website builder does not show the engineering depth a product company cares about.

04

Token open-source contributions

A single pull request fixing a typo will not help. Months of meaningful contributions can be a capstone in itself, but drive-by commits done to tick a box are easy to spot.

Finding your idea: start with yourself

The question I get most is "what project should I build?" People want a neat answer, but if I hand you an idea and you implement it, you are back to following instructions. The whole point is to flex your own problem solving and creativity.

Naval Ravikant says you escape competition through authenticity: nobody can compete with you at being you. Your background, interests and frustrations are fertile ground for ideas nobody else would think of. A few places to look:

Problems you have personally faced. If you worked in healthcare, maybe clinic appointment systems were a nightmare. In finance, maybe you saw inefficiencies in budgeting or compliance that a small tool could fix. Your industry knowledge shows you pain other people miss. What you are passionate about. A fitness enthusiast could build workout challenges for gym friends; a musician could build a way for friends to collaborate on playlists. When it is personal you stay motivated, and the passion shows.

Manual processes you could automate. Maybe a friend's small business spends hours scheduling posts or tracking inventory in Excel. A modest tool that clearly saves time or money is impressive, and you instantly have a real user giving feedback. Tools for developers. Notice the moments you think "this is so tedious" while coding. A VS Code extension or a script that writes better commit messages shows advanced thinking, and your test users are all around you.

Train the idea muscle

The first time you sit down to brainstorm, you might draw a blank. That is fine. Write down every frustration or inefficiency you have hit at work, school or home, no matter how small. "My friend's shop has trouble tracking deliveries." "My study notes always got disorganised." "My gardening club needs a way to coordinate tasks." Any of those could be the seed of a project.

Your capstone does not have to sound sexy. It has to be useful. Jason Fried, the founder of Basecamp, has a philosophy I admire: make something unique and make it well. Version one will not be everything to everyone, but it can be exactly what you need and nothing more. Solve your own problem deeply, even for a tiny audience. That focus will stand out far more than a trendy app you do not care about. If you need a starting point, browse these coding projects for beginners and then make one your own.

Building it like a real product

A big mistake is cramming in every new technology or being wildly over-ambitious. The project is first and foremost for you to learn and demonstrate skill. Users and revenue are a bonus. Pick a stack aligned with the jobs you want and go deep. If you have been learning JavaScript, React on the front end, Node or Express on the back end and MongoDB or Postgres for data is a popular, in-demand combination. As a self-taught developer, I doubled down on marketable tools I could master instead of chasing shiny frameworks. Experienced people would rather see depth in one stack than a list of twenty languages.

Plan your core features like a real product. For a gardening club app: login, a task list for chores, a meetup calendar and a harvest log. That is already a lot, so decide the order and think in terms of an MVP, the smallest version that is still useful. Nobody gives you a step-by-step guide here, which is exactly the real-world skill you are practising.

Use good habits from day one. Initialise a Git repository and commit regularly; it saves you when you break things and shows anyone looking how you work. Write a clear README explaining what it is, what tech you used and how to run it. And deploy it. A hiring manager with five minutes will click a link and play with your app; they will not clone and compile your repo. For a mobile app, record a short demo video. Treating your project seriously signals you will treat theirs seriously too.

Get stuck, ship early, iterate

You will hit roadblocks: a slow database query, UI state that goes wonky. That is where the real learning happens. Debug systematically, read the documentation and ask on forums. Most junior developers stop when it gets hard. If you keep going, that is how you set yourself apart.

Once you have a version one, even a basic one, get feedback. Ask your target users, former colleagues or a relevant online community to try it and be honest. Even a handful of users is gold. Incorporating feedback and shipping updates shows you can improve a product over time, which is advanced stuff for a junior developer.

When we first launched Sigmaschool, we tested the concept with a minimal offering and about RM 3,000 of test ads, and the response was so strong it was clear we were solving a real need before the product was even fully built. Putting something out there, imperfect, gave us insights we would never have got by staying in development mode. Your capstone works the same way. Do not hide it until it is perfect. And if someone ends up using it every week or paying to cover your server costs, great, but keep your eyes on the main prize: learning and showcasing skills.

“The difference between a dabbler and a developer is finishing.”

Deric Yee, in the video

The mindset: grit, and think like an owner

Building a capstone is not easy. If it were, every junior developer would have one. Most start something ambitious, hit a snag, lose interest and let it fade. Push through those moments and you have already separated yourself from the pack. Do not settle at "good enough to maybe get a job". You grow most when you stretch beyond your comfort zone.

Think like an owner, not just an employee. Builders like Andrew Wilkinson and Pieter Levels inspire me here. Treat your project as your own product: care about quality, the user's perspective and scope, and wear several hats, developer, tester, even marketer. In an interview, saying "I really cared about how users would experience this" or "I had to work out hosting costs so it could run cheaply" makes you sound mature and responsible. Document the journey on LinkedIn or X too; people have reached out to me because of posts about things I built.

What about AI?

Will AI just code everything, making all this pointless? AI is changing the developer's role, but it is not replacing human creativity and problem solving. With AI automating routine tasks, those skills are more valuable than ever. As I told a journalist recently, AI is widening the gap between engineers who know how to leverage these tools and those who do not. Be in the first camp. Use AI to brainstorm and debug, or build an AI API into your capstone itself. Our guide on how to learn coding with AI covers how to do that without becoming dependent.

So here is the challenge: what would your capstone be? It might start as a weekend hack for a friend. Nurture it and use it to learn the deeper stuff. In an interview, "I know React" becomes "I built a React and Node app that helped 50 volunteers coordinate relief efforts, and learned to optimise an interactive map along the way". That is night and day.

If you want structured support to get there, our AI-Native Software Development Programme is built around hands-on projects and real-world skills, with AI woven through, and comes with a money-back guarantee. If you want to see how we teach first, start with the free trial. Go and build something only you could build, and let it be your calling card.

Questions people ask

  • What is a capstone project for a junior developer?

    A capstone project is one substantial, end-to-end application you design and build yourself to solve a real problem. It usually covers front end, back end, a database and deployment. Unlike tutorial projects, the idea, feature choices and problem solving are yours, so it gives employers direct evidence that you can take an unclear problem and turn it into working software.

  • How many portfolio projects do I need?

    Fewer than most people think. One well-crafted capstone that you can explain in depth usually carries more weight than a dozen small tutorial apps. Small practice projects are useful while you learn the basics, but for job applications, depth, originality and evidence of real users or feedback matter far more than the number of repositories on your GitHub.

  • How do I come up with a capstone project idea?

    Start from your own life. List problems you faced in past jobs, frustrations in daily life, hobbies you care about, manual processes friends or family repeat every week, and tedious moments you hit while coding. Any of these can become a project. Aim for something genuinely useful to a small group of people rather than a flashy idea you have no connection to.

  • Should I deploy my capstone project?

    Yes, wherever possible. A live link lets a busy hiring manager try your app in a minute, while asking them to clone and run your code means most never will. There are free and low-cost hosting options for web apps, and for mobile apps a short demo video works well. Pair the deployment with a clear README on GitHub.

Your next step

Our 12-week AI-Native Software Development Programme is a live, mentor-reviewed cohort where you build and ship real products with AI, and learn to explain every line. Start with the free trial to see how we teach, then decide.

Watch next

All videos

Learn the fundamentals that pay off in any direction.
Coding plus AI, taught properly.

Whatever you end up doing, the people who understand how software and AI actually work get to direct them. Try the first lessons free, no card needed, or see the full 12-week AI-Native Software Development Programme.