Career prep
Portfolio and Projects Guide — Prove You Can Do the Job
Certificates say you studied. Projects show you can build. This guide covers which projects to build for your target role and how to present them so a reviewer understands them in two minutes.
Est. 15 min November 2026
Two or three strong projects beat ten tutorials
Reviewers rarely look at more than two projects. A small number of complete, well-explained projects that look like real work will do more for you than a long list of tutorial clones. Aim for:
- One flagship project — end to end, deployed or runnable, close to the job you want. Your career path capstone is designed for this.
- One or two supporting projects — each showing a different skill (for example streaming, evaluation, infrastructure as code, or a polished UI).
What to build for your role
- AI / Generative AI engineer — a RAG assistant with citations and an evaluation set; a tool-using agent with human approval. See the RAG capstone and agent capstone.
- Data engineer — a batch plus streaming pipeline with orchestration, data quality checks, and a dashboard. See the data engineering capstone.
- Data analyst / scientist — an analysis of a real public dataset with a clear question, SQL, a model or experiment, and a written recommendation.
- Full-stack / backend / frontend — a deployed app with authentication, a database, tests, and CI. See the full-stack capstone.
- Cloud / DevOps / SRE — an app deployed with Terraform, containers, CI/CD, and monitoring dashboards. See the DevOps capstone.
- Security — a home lab write-up with detections, or a red-team report on a sample LLM app.
The README is the project
Most reviewers read the README and never open the code. Use this outline:
- One-line summary — what it does and for whom.
- Demo — live link, screenshots, or a 2-minute video.
- Architecture — a diagram and a short explanation of the data or request flow.
- Tech stack and why — the choices you made and the alternatives you rejected.
- How to run it — copy-paste setup steps that actually work.
- Results — numbers: latency, accuracy, rows processed, cost, test coverage.
- What I would do next — honest limits and improvements. This shows engineering judgement.
GitHub hygiene
- Pin your best repositories on your profile and write a short profile README.
- Never commit secrets. Use environment variables and a
.env.examplefile. - Meaningful commit messages and a clean folder structure — reviewers notice.
- Add basic tests and a CI workflow so the green check mark is visible.
Show your work in public
- Write a short post about one hard problem you solved in each project.
- Share a demo video on LinkedIn with what you learned — not just “I built X”.
- Contribute small fixes to open-source tools you use; they make strong resume bullets.
Next steps
Put your projects on your resume, then practise explaining each one in three minutes using the mock interview guide.
Share
Discussion
0Have a better approach? Found something outdated? Share it — your knowledge helps everyone learning here.
Loading...
