Role Up

Role Up App Page

Inspiration

The inspiration for Role Up app came from my own struggle of switching job into a different domain. I barely had any idea about essential skills required for my desired role and from where to start preparing for it. I had to look into various sources to get a basic roadmap and browsed through multiple courses that could make me competent for that role. All I wished for was a one stop solution where all this information was available on single click.

What it does

The Role Up App provides a roadmap for any desired role along with some of the best courses available online that you can take up to gain the required skillset. It also provide some role related interview questions to help you prepare for the job and some job opening links as well based on location and years of experience. It also features AI assistant that can help you with some resume samples, provide answer to interview questions, more course links and basically anything you ask related to the role.

How I built it

This app was built on PartyRock, the AWS Bedrock playground using generative AI and absolutely zero line of code. Structure is very simple, it has User Input widget to get role, years of experience and location from user and text generation widget which has been configured by giving a simple prompt to generate a roadmap for the role entered . Similar action taken for other text widgets while for the AI assistant Chatbot widget is used.

Challenges I ran into

The biggest challenge I ran into was making course links clickable. Initially it was generating links in simple text format that user had to manually copy and paste in browser. This was degrading user experience. It took multiple changes in prompt to make hyperlinks clickable. Even now it doesn't work as expected sometimes :(

Another challenge was generating job links or course links that user can actively apply to but it seems out of scope of the FMs as it threw error that the LLM can only decipher text.

Accomplishments that I am proud of

Again figuring out prompt to convert link from text format to clickable format was my notable accomplishment in this project. It took most of my time, otherwise the complete project was made in just an hour. Also the fact that it solves a real world problem including the one of my own really excites me. Without PartyRock, I couldn't have imagined making this app and even if I did, definitely wouldn't have imagined in my wildest dream that I will be able to create something like this in just few hours!

What I learned

With the help of this Project I learnt -

How Gen AI and prompt Engineering works and how AWS PartyRock has harnessed these technologies to enable people create apps without using even single line of code.
Foundation Models (FMs) and its features.
Different Models that AWS PartyRock (and Bedrock) offers like Claude, Jurassic etc. and how choice of model affects the output.