Inspiration
Our original idea was to build a University of Toronto (UFT) alumni network using the LinkedIn API. We wanted to help students connect with alumni to find career opportunities. However, we quickly hit a roadblock—LinkedIn’s API was extremely limited, providing only basic profile information (name, email, and image) and no search functionality. We stayed up until 2 AM on Friday night trying to make it work but ultimately had to pivot.
The next morning, inspired by our frustration and our shared passion for career growth, we came up with a new idea:
What if we could use your resume to map your future career path and show you how to reach your dream job? That’s how RoadMapr was born.
What it does
RoadMapr is an AI-powered career planning tool that:
Analyzes your resume to understand your skills, experiences, and career trajectory.
Generates a personalized career roadmap, showing potential roles, companies, and skill gaps.
Provides actionable insights with company rationales to help you choose the right path.
Stores your career roadmap for future updates and adjustments.
You simply upload your resume, and RoadMapr uses LLMs to map out your journey step-by-step.
How we built it
- Backend: Built with Python, using Cohere's API to generate career roadmaps from resumes.
- Database: We used Supabase to store resumes, career roadmaps, and user profiles.
- Frontend: Developed with React and Tailwind CSS, providing an interactive experience with career phases and animations.
- LLM Integration: We refined prompts to improve the quality and relevance of the career roadmaps.
Challenges we ran into
Pivoting from our original idea: The LinkedIn API’s limitations were a huge disappointment, forcing us to scrap our original project after hours of work.
Prompt Engineering: Refining the prompts for the LLM was tricky. We needed the AI to output structured JSON with clear career phases, which required multiple iterations.
Frontend Animations: We faced issues with rendering smooth animations for the roadmap phases but resolved them after experimenting with different libraries.
Time Crunch: Losing almost a full day to the LinkedIn API setback meant we had to work under extreme pressure to finish the project.
Accomplishments that we're proud of
Creative Pivot: Turning a major setback into a new, exciting project overnight.
LLM Integration: Successfully generating structured career roadmaps with company rationales.
Frontend Animations: Creating a smooth, interactive roadmap experience.
Teamwork & Resilience: Pushing through challenges and working together to ship a fully functional MVP.
What we learned
Adaptability: Sometimes, the best ideas come from setbacks.
Prompt Engineering: Getting structured, high-quality outputs from LLMs requires experimentation and iteration.
Time Management: Rapidly building an MVP under pressure taught us the importance of prioritizing features.
API Limitations: It’s crucial to research API capabilities deeply before building around them.
What's next for RoadMapr
Resume Parsing Improvements: Use advanced NLP to extract skills and achievements from resumes more accurately.
Interactive Roadmaps: Allow users to customize their career paths by adding or removing roles.
Skill Gap Analysis: Suggest certifications, courses, and projects to help users reach their next career goal.
Company Insights: Provide real-time insights into company hiring trends using web scraping and job boards.
Alumni Network (Long-Term): Revisit our original idea using verified alumni databases and other networking APIs.
RoadMapr: Your roadmap, your way.
Built With
- cohere
- python
- react
- supabase
- typescript
Log in or sign up for Devpost to join the conversation.