Inspiration
The traditional format of education—hours of videos or long-form courses—doesn't align with the reality of today's fast-paced world. We noticed that many people, whether students or professionals, have strong ambitions but struggle to find consistent blocks of time to learn. We built Skill Bridge to solve these two problems: to make learning highly adaptable to any schedule (even just 5 minutes) and to provide clear, actionable career roadmaps based on a user's actual resume.
What it does
Skill Bridge is an AI-powered, personalized learning platform featuring three core experiences:
Quick Learner: Users input a topic and their available time (e.g., 5 to 60 minutes). Our AI generates a concise, bite-sized summary tailored exactly to that time limit. It includes an interactive Q&A chat where users can ask follow-up questions, followed by dynamically generated quizzes to test their retention, producing a final performance report.
Career Trajectory Analysis: Users upload their resumes (PDF/DOCX), and our system uses NLP to extract their skills, education, and experience. It then analyzes this data, suggests optimal target roles, identifies critical skill gaps, and generates a personalized "Career Roadmap" linking them to actual courses to bridge those gaps.
Dynamic User Profiles: The platform maintains real-time user profiles that can be instantly updated or re-synced simply by re-uploading an updated resume, keeping the learning journey aligned with their career progress.
How we built it
We built Skill Bridge focusing on a seamless user experience backed by powerful data processing:
Backend: We utilized Python with the FastAPI framework to build high-performance, asynchronous RESTful APIs.
AI Integration: The core intelligence of the platform is powered by the Google Gemini API, which handles summarizing vast topics into strictly timed segments, powering our conversational Q&A agent, generating quizzes, and performing the complex NLP required for our resume analysis.
Data Processing: We used tools like PyPDF2 and python-docx to handle resume parsing directly on the server before passing the raw text to our AI pipelines. Data persistence (user profiles, auth) is handled via a connected database strategy (MongoDB).
Frontend: We designed a fully responsive, modern UI using HTML, CSS (with Glassmorphism and dark/light modes), and vanilla JavaScript.
Challenges we ran into
One of the primary challenges was hallucination mitigation and content scaling with the AI. Prompt engineering was tricky: ensuring the Gemini API generated a 5-minute read when asked for 5 minutes, but a comprehensive 30-minute guide when asked for 30 minutes, required extensive prompt tuning and validation checks. Another major hurdle was Resume Parsing Accuracy. Real-world resumes have wild formatting; standardizing text extraction across various PDFs and DOCX files before feeding them into our AI pipeline for career analysis took significant trial and error.
Accomplishments that we're proud of
We are incredibly proud of the Quick Learner feedback loop. We successfully connected together an AI summary generation, live Q&A chat memory, and a dynamic quiz evaluation system into one seamless user session that actually evaluates user performance in real-time. We're also very proud of the Resume-to-Roadmap pipeline—turning a static, cluttered PDF into a structured JSON profile and actionable career advice in under a few seconds feels like magic.
What we learned
We learned a tremendous amount about Prompt Engineering and AI Context Windows.
What's next for Skill Bridge
To drive continuous engagement and motivation, we would like to introduce a gamification engine. Users will be able to track learning streaks, earn skill-specific badges, and unlock milestone rewards to make upskilling as fun as it is rewarding.
Log in or sign up for Devpost to join the conversation.