Pathfinder AI

An AI career companion that helps anyone discover their passions and own their path — through real conversation, not forms. Powered by Amazon Nova.

Comment

Inspiration

What it does

How we built it ## Inspiration

In my community, a lot of kids don't know what's out there. They carry pressures no kid should — working to earn a living when they should be discovering their passions. It's a broken cycle. People end up doing work they don't enjoy, not by choice, but to survive.

The little programs that exist don't have tools people can continue to use long after the program ends. I was lucky — I found my passion early. But most people don't get that chance. PathFinder AI is built to change that.

What it does

PathFinder AI is an AI-powered career discovery and development companion. It guides users from discovering their passions to landing their dream job — all through natural conversation powered by Amazon Nova.

  • Discovery — Real conversation to uncover passions and career aspirations. Not a form, not a quiz.
  • Certificate Analysis — Upload any cert image or PDF and Nova extracts credentials automatically using multimodal understanding.
  • Profile & CV Generation — Nova synthesizes everything you've shared into a professional profile and tailored CV.
  • Journey Tracking — Goals, achievements, and challenges auto-detected from conversation.
  • Job Matching — Real job listings matched to your skills and direction.
  • Any field — Tech, sports, trades, beauty, music, healthcare — whatever drives you.

How we built it

The entire backend runs on AWS. Amazon Nova Lite powers every AI interaction through the Bedrock Converse API — the discovery conversation, certificate analysis, profile generation, and CV creation. Six AWS Lambda functions handle different domains (messaging, state, portfolio, journey, opportunities, admin). API Gateway with Cognito authentication secures the API. User data lives in S3 and DynamoDB. The React/TypeScript frontend is hosted on S3 behind CloudFront.

Challenges we ran into

  • Getting Nova to reliably extract structured data from multi-turn conversations while maintaining natural flow
  • Multimodal certificate analysis — handling varied image quality, PDF formats, and different certificate layouts
  • Designing a conversation engine that transitions smoothly between four phases (Discovery → Skills → Path → Coaching) without feeling robotic
  • Building journey auto-tracking that detects goals, challenges, and achievements from natural language without explicit user input

Accomplishments that we're proud of

  • A working end-to-end career companion that takes someone from "I don't know what I want to do" to having a professional profile, CV, and matched job listings
  • Multimodal cert upload that actually works — snap a photo of any certificate and Nova reads it
  • The conversation feels like talking to a real mentor, not filling out a form
  • Built for any field, not just tech — because career guidance should be universal

What we learned

  • Amazon Nova Lite is remarkably capable at maintaining context across long multi-turn conversations
  • The Converse API makes it straightforward to build structured AI interactions with tool use
  • Multimodal understanding opens up possibilities that text-only models can't touch — cert analysis was a game changer
  • Building for communities that lack career guidance requires a fundamentally different approach than traditional career tools

What's next for PathFinder AI

  • WhatsApp integration to meet users where they are
  • Universal job matching across all industries (not just tech)
  • Real-time course APIs for verified learning resources
  • Local events and workshops via Eventbrite integration
  • Multi-language support
  • Mentor matching to connect users with real people in their field
  • Interview prep

Built With

Share this project:

Updates