Pathly

🌱 Pathly: AI Career Coach

Inspiration

As recent graduates, we’ve all had that same realization —

“If only I had known what to learn earlier, I could’ve been better prepared.”

University life often leaves students unsure about what skills to build next or how to align their studies with real-world career paths.
Curricula can be outdated, information is scattered, and guidance often feels one-size-fits-all.

We wanted to change that.
Pathly was born from our wish to give every student the kind of personalized, data-driven career guidance we wish we had — powered by AI and built on AWS.

What it does

Pathly acts as an AI Career Coach that bridges the gap between academia and industry.

It helps students:

Plan their academic and career paths from university to post-graduation
Track their skills and competencies over time
Get AI-powered learning and course recommendations tailored to their goals
Receive peer reviews that feed into AI-driven soft skill assessments
Discover industry trends and close their skill gaps with real, actionable steps

Using Amazon Bedrock, Pathly analyzes each student’s profile, peer feedback, and job market data to generate a personalized roadmap — guiding them on what to learn next to stay relevant and confident.

How we built it

Our architecture is fully deployed on AWS for scalability and reliability.

Frontend: Built as a web app hosted on Amazon EC2, providing a responsive and intuitive interface.
AI Engine:
- Powered by Amazon Bedrock, which performs reasoning and generates personalized learning paths.
- AWS Lambda connects Bedrock to our PostgreSQL database for secure data retrieval and updates.
- Bedrock also integrates with a web crawler to continuously analyze job postings and extract trending skill requirements.
Database:
- PostgreSQL stores user data, skills, and progress metrics.
Storage & Integration:
- Peer reviews, academic progress, certificates and portfolios feed into Bedrock for holistic insights.

Together, these services form a scalable, reasoning-capable system that evolves with each student’s learning journey.

Challenges we ran into

Designing seamless data flow between Bedrock, Lambda, and PostgreSQL while maintaining security and low latency.
Crafting prompts that produce trustworthy, specific, and actionable AI recommendations.
Handling large and diverse job market data from our web crawler and filtering only relevant insights.
Ensuring the UX stayed simple and friendly for students — avoiding “tech overload” while still offering powerful features.

Accomplishments that we're proud of

Built a fully functional AI agent on AWS Bedrock capable of reasoning across academic and job data.
Designed an end-to-end personalized learning platform that integrates both hard and soft skill insights.
Created a feedback ecosystem where peer reviews directly enhance AI recommendations.
Delivered a working prototype that demonstrates how generative AI can transform education — not by replacing teachers, but by empowering students.

What we learned

This project taught us that building impactful AI tools isn’t just about algorithms — it’s about empathy, context, and clarity.

We learned how to:

Architect scalable, cloud-native systems with Amazon Bedrock and AWS Lambda.
Design multi-source reasoning workflows that combine structured data, peer input, and live market intelligence.
Use AI ethically and responsibly in educational settings.
Collaborate effectively as a team across technical, design, and product roles.

What's next for Pathly

Our next step is to partner with our university to pilot Pathly with real students.
We want to help them plan their courses, track progress, and get AI-backed guidance on the skills that matter most for their careers.

Looking ahead, we plan to:

Expand to more universities and integrate with academic systems.
Add Amazon Bedrock AgentCore for multi-agent collaboration and reasoning.
Enhance predictive analytics for career trajectory modeling.
Integrate Amazon Q to deliver interactive, real-time tutoring and course insights.