Launchpad AI

LaunchPad AI Architecture Diagram

Inspiration

Students and young professionals often face confusion when planning their careers. They’re unsure what roles fit them, what skills they need, and where to start. This inspired us to create LaunchPad.AI — an autonomous AI career copilot built on AWS that brings personalized, data-driven career guidance to everyone.

What it does

LaunchPad.AI helps users plan, prepare, and apply for jobs — completely autonomously.

Analyzes user profiles, resumes, and goals using Amazon Bedrock and SageMaker.
Identifies relevant career paths and detects key skill gaps.
Recommends tailored courses and certifications to upskill effectively.
Generates personalized resumes, cover letters, and LinkedIn summaries.
Tracks job applications, deadlines, and even sends automated reminders or follow-ups.

LaunchPad.AI acts as a full-fledged AI career assistant that adapts to each user’s journey.

How we built it

We designed LaunchPad.AI using a modular AWS architecture:

Amazon Bedrock (Claude / Nova) for reasoning, dialogue, and goal understanding.
Amazon SageMaker for user profiling, skill-gap analysis, and recommendation logic.
*AWS Lambda * for orchestration, task automation.
Amazon Bedrock AgentCore to enable autonomous task execution and tool calling.
Amazon DynamoDB / S3 for user memory, progress tracking, and file storage.
Amazon Cognito for authentication and access control.
AWS Amplify (React.js) for front-end deployment and user interface.
Python (Boto3, LangChain) for backend intelligence and workflow logic.

Mathematically, the system ranks career paths using a weighted model:

$$ C_i = \alpha S_i + \beta G_i + \gamma L_i $$

where
( S_i ) = skill match,
( G_i ) = goal alignment,
( L_i ) = learning path relevance,
and ( \alpha, \beta, \gamma ) are adaptive weights optimized via SageMaker feedback loops.

Challenges we ran into

Integrating multiple AWS services smoothly through AgentCore primitives.
Managing LLM context persistence with DynamoDB and S3.
Balancing autonomous decision-making with safe and ethical AI behavior.
Ensuring low-latency responses while maintaining reasoning quality.
Designing a user experience that felt both intuitive and powerful.

Accomplishments that we're proud of

Built a working autonomous AI agent that performs multi-step reasoning and tool execution.
Deployed a live prototype hosted on AWS Amplify with seamless backend integration.
Achieved contextual memory and self-updating progress tracking for each user.
Designed a scalable architecture fully compliant with AWS best practices.
Developed an end-to-end demonstration video showcasing intelligent reasoning and action chaining.

What we learned

How to design and deploy agentic workflows using AWS Bedrock AgentCore.
The importance of structuring reasoning + autonomy + integration to create real value.
LLMs can excel in career coaching and education, when combined with personalized data.
Balancing human control with AI autonomy is key to building user trust.
We learned to think like “builders of systems that think” — not just app developers.