AWS Builder Center

Builder Center

7 AWS + AI Projects Every Student Should Build (With Prompts to Start Each One)

7 AWS + AI Projects Every Student Should Build (With Prompts to Start Each One)

Seven beginner-friendly cloud and generative AI project ideas that make a strong portfolio, each with a ready-to-use Claude Code prompt, the AWS services involved, and a cost-saving tip. Pick one and ship it this week.

Most students collect certificates. Hiring managers click on projects.
If you want something real to show on your resume, here are seven AWS + AI projects you can build as a student, each small enough to finish in a weekend. Every one comes with the services you'll use, a starter prompt for Claude Code, and a way to stretch it into something more impressive.
Rule of thumb: a deployed project with a README and a diagram beats ten half-finished tutorials.

Before you start (2 minutes)

  • Set an AWS Budget alert so you never get a surprise bill.
  • Use least-privilege IAM permissions, never root credentials.
  • Keep secrets out of your code and out of Git.
  • Delete resources you're not using.

1. Study Assistant API

What it does: answers questions in simple language and ends with a practice question. Services: Amazon Bedrock, AWS Lambda, API Gateway Starter prompt:
Plain text
Plan, then build, a Python Lambda function that takes a question from an

API Gateway POST request and returns a short explanation using Amazon

Bedrock's Converse API. Add input validation, a max output token limit,

and unit tests.
Level up: save question history in DynamoDB.

2. Notes Summarizer

What it does: upload a text file, get a summary with key points. Services: Amazon S3, Lambda, Bedrock Starter prompt:
Plain text
Create a Lambda function triggered by new .txt uploads to an S3 bucket.

It should summarize the file with Bedrock and save the summary to a

second bucket. Include error handling and tests.
Level up: add an email notification with Amazon SNS.

3. Serverless Portfolio Website

What it does: hosts your personal site (and links to these projects). Services: S3, CloudFront, Route 53 (optional) Starter prompt:
Plain text
Help me write a simple static portfolio site in HTML/CSS, and give me

step-by-step instructions to host it on S3 behind CloudFront with HTTPS.
Level up: automate deployment with GitHub Actions.

4. Quiz Generator

What it does: turns any topic into 5 multiple-choice questions. Services: Bedrock, Lambda, API Gateway Starter prompt:
Plain text
Build a Lambda function that asks Bedrock for 5 multiple-choice questions

on a given topic and returns them as validated JSON. Add a retry if the

JSON is malformed, and write tests for the parser.
Level up: add a leaderboard with DynamoDB.

5. Expense Tracker with Receipt Reading

What it does: upload a receipt image, extract the amount and category. Services: S3, Lambda, Amazon Textract, DynamoDB Starter prompt:
Plain text
Create a pipeline where an uploaded receipt image in S3 triggers Lambda,

which uses Textract to extract totals and stores a record in DynamoDB.

Include IAM permissions that follow least privilege.
Level up: add a monthly summary report.

6. Resume Feedback Bot

What it does: gives structured feedback on a resume (clarity, impact, gaps). Services: Bedrock, Lambda, S3 Starter prompt:
Plain text
Write a Lambda function that takes resume text and returns feedback in

three sections: strengths, weaknesses, and suggested rewrites. Do not

store the resume anywhere, and make sure nothing sensitive is logged.
Level up: compare a resume against a job description.

7. Cloud Cost Watchdog

What it does: checks your spending daily and warns you before it grows. Services: AWS Budgets, EventBridge, Lambda, SNS Starter prompt:
Plain text
Create a scheduled Lambda that runs daily via EventBridge, reads my

cost data, and sends me an SNS email if spending passes a threshold.

Explain every IAM permission it needs.
Level up: add a weekly cost summary written by Bedrock.

How to turn any project into a portfolio piece

  1. Draw the architecture (even a simple diagram helps).
  1. Write a clear README: problem, solution, how to run it, cost notes.
  1. Add screenshots or a short demo.
  1. Publish it on GitHub and write about it here on Builder Center.
  1. Explain what went wrong and how you fixed it. That shows real learning.

Using Claude Code well

Use the same loop for every project: plan first, build in small tasks, run the tests, and review every diff before committing. The agent speeds you up, but understanding the code is still your job, and interviewers will ask about it.

Your turn

Pick one project and finish it this week. Not all seven, just one.
Which one are you building? Tell me in the comments, and I'll check back and share tips.

Resources

# agentic-ai# generative-ai# amazon-bedrock# students# serverless
Any opinions in this article are those of the individual author and may not reflect the opinions of AWS.
Enjoyed reading this content? Let the author know!

Your likes, comments, shares, and saves help creators reach more builders.

Loading recommendations

Loading article