SkillMatchAI

Logo

     What it does

SkillMatchAI is your personalized career co-pilot. You upload a resume, answer a few tailored prompts, and our AI identifies your current strengths and missing skills. Then, it recommends relevant jobs along with personalized learning paths—like Udemy courses or degree programs—to help bridge the skill gap and reach your goals, all within your budget and timeline.

    How we built it

We used:

React + Tailwind for a clean frontend UI
Flask backend to handle API routes and logic
Gemini API for NLP-driven resume parsing, questions generation, and skill mapping
Pinecone for vector-based job, course, and program recommendations
MongoDB to persist user profiles and session data
Firecrawl + Gemini for scraping and structuring real course data from the web
```
Challenges we ran into
```
Structuring skill domains to balance granularity and generalization
Ensuring accurate and budget-aware course recommendations
Preventing hallucinations and duplicate recommendations from AI
Building a clean UI that smoothly transitions across multiple steps

Accomplishments that we're proud of
Built a full-stack, AI-powered, personalized skill recommender from scratch in <36 hours
Integrated Pinecone vector search across three datasets
Created a natural UX flow that guides users from resume to personalized action steps
Used Gemini in both RAG and agentic formats (scraping + summarization + classification)

What we learned
Prompt design is everything when working with generative AI
How to handle fallback logic when vector search returns low scores
Better ways to guide users through survey UIs and track their intent
How to align user goals with tangible, data-driven paths using AI

What’s next for SkillMatchAI
Allow users to track progress over time and re-evaluate goals
Add LinkedIn and GitHub integration for even richer profiles
Expand to soft-skill and leadership training recommendations
Collaborate with educational platforms for deeper API-level course access
Fine-tune Gemini prompts using real user data and feedback