About Career Restart AI

Career Restart AI was inspired by a real and growing problem — millions of professionals, especially women and career-break candidates, struggle to re-enter the workforce after time away from their careers. Having personally experienced a career break, I wanted to build something meaningful that could provide clarity, confidence, and a practical roadmap for people restarting their careers.

Most job seekers face three major challenges:

They don’t know which skills they are missing

They don’t know how to bridge the gap

They don’t know where to start

Career Restart AI was designed to solve exactly this using Amazon Nova’s reasoning capabilities.

How I built the project

I built Career Restart AI as a web-based Agentic AI application using:

Amazon Nova (via AWS Bedrock) for reasoning and analysis

Stream lit for the user interface

Python for backend logic

The user uploads their resume and pastes a job description. The system then sends both to Amazon Nova 2 Lite, which acts as a career-coaching agent. Nova reasons about the candidate’s background, compares it with the job requirements, and generates:

Missing skills

Learning recommendations

A personalized career roadmap

This makes the AI behave like a career coach agent, not just a text generator.

What I learned

Through this project, I learned:

How to build agentic workflows using Amazon Nova

How to connect AI reasoning models with a real user interface

How to design AI around human problems, not just technology

I also learned how to turn a real-life problem into a working AI product that can be demonstrated, tested, and used.

  • Challenges I faced

Some of the main challenges included:

Setting up AWS Bedrock and Nova model access

Connecting cloud AI models to a local Stream lit app

Designing prompts that made Nova behave like a career coach

I overcame these by breaking the problem into small steps — first building the UI, then connecting Nova, and finally refining how the AI analyzes resumes and job descriptions.

Built With

Share this project:

Updates