![MyCareerPath- A Creative Assistant](https://assets.community.aws/a/2dYQpILUmRUDqgUBlQX030dQ9ds/Scre.webp)
MyCareerPath- A Creative Assistant
Using Generative AI to prepare roadmaps for interview preparation and gather study material to master any topic!
Published Mar 11, 2024
As a working professional I often struggle to prepare a timetable and gather relevant resources for interview preparation and self study. My struggles and a quote by the Prime minister of India, “Turn disaster to opportunity” (translated) inspired me to develop an application to address the concern.
MyCareerPath is a Generative AI application that accepts user input on the subject they want to learn (example- Machine Learning), the amount of time they want to dedicate (in weeks, months, or days), and their preferred resources (YouTube videos, blogs, or books). The app uses the Claude LLM to generate a study plan as well as MCQ quizzes for self assessment or interview preparation.
The app presents a concise and structured roadmap to prepare any topic as per time availability. It collects all study resources and quizzes from across the web and mentions them at one place including their URLs. Therefore, the solution can be presented to working professions worldwide for their interview preparation, self-study, or preparing for certifications. The provided URLs seldom contain very useful and informative material which can be shared and bookmarked for reference.
In case the AWS PartyRock platform was unavailable, I would propose an architecture of three AWS services- API Gateway, AWS Lambda, and AWS Bedrock (illustrated below). The core functionality revolves around exposing an API endpoint via API Gateway to accept user input.
![](https://assets.community.aws/a/2dYTMFT2fgq4qAU44ELL82oLZ2b/Scre.webp)
Purpose of each service-
- API Gateway: receive the user input (preparation time, preferred resources, and target course) and the prompts for the LLM
- Lambda: invokes a LLM (anthropic Claude) hosted on AWS Bedrock
- Bedrock: generates text responses using the prompt and user inputs provided
Team members:
Akshay Vayak
Divyank Singh