ai-job-tracker
AI-Powered Job Tracker with Smart Matching
A full-stack AI application that parses resumes, fetches real-time jobs, and ranks them using AI-based matching. The system includes filtering, top matches, and an AI assistant to control job preferences.
Why This Project Matters
Job searching is time-consuming and often involves manually filtering hundreds of listings. This project uses AI-powered semantic matching and workflow automation to help candidates discover relevant opportunities faster and more accurately.
🚀 Live Demo
👉 https://zippy-quokka-0cd1c3.netlify.app/
📂 GitHub Repository
👉 https://github.com/DSatyakeerthi/ai-job-tracker
🔥 Key Features
- AI-powered semantic job matching
- Resume parsing using LLMs
- Embedding-based similarity scoring
- Fastify backend APIs
- Real-time job filtering
- Vector search workflows
- Authentication and user workflows
🏗️ System Architecture
👉 
Flow: Resume → Parsing → Embedding → Job Fetch → Matching → Ranking → UI Display
🛠️ Tech Stack
Frontend:
- React
- JavaScript
- HTML/CSS
Backend:
- Fastify (Node.js)
- REST APIs
AI / ML:
- OpenAI API
- RAG (Retrieval-Augmented Generation)
- Embeddings
Data & Processing:
- JSON storage / In-memory processing
- Job API integration (Adzuna)
Tools:
- Windsurf (AI-assisted development)
- GitHub
Backend APIs
- Resume Upload API
- Job Fetch API
- Semantic Matching API
- Recommendation API
- Authentication API
Project Structure
frontend/ → React frontend
backend/ → Fastify backend APIs
docs/ → Screenshots & architecture
data/ → Resume and job data
📸 Screenshots
Home / Job Listings
Filtering (24 hours / 1 week)
🧠 How AI is Used
- Resume is uploaded and processed using LLM APIs
- Extracted data is structured into skills and experience
- Jobs are fetched from external APIs
- Matching is performed using:
- Embeddings similarity
- Rule-based scoring
- LLM-assisted evaluation
- Top matches are ranked and displayed to the user
Challenges Faced
- Handling inconsistent resume formats
- Improving semantic matching accuracy
- Managing API response latency
- Designing scalable workflow pipelines
- Balancing rule-based and embedding-based scoring
💡 Key Learnings
- Designing AI systems beyond simple API calls
- Handling large data using chunking and pipelines
- Improving AI output reliability using validation logic
- Building end-to-end applications combining AI + backend + UI
🚧 Future Improvements
- Add database (PostgreSQL / MongoDB)
- Improve UI/UX for better user experience
- Enhance matching accuracy with advanced ranking models
- Deploy at scale with cloud infrastructure (AWS / Render)
My Contributions
- Built backend APIs using Fastify
- Integrated OpenAI APIs and embeddings
- Developed semantic matching workflows
- Implemented resume parsing pipelines
- Designed filtering and recommendation logic
- Improved debugging and workflow automation
Deployment Architecture
Frontend: Netlify
Backend APIs: Render
AI Services: OpenAI API
External Data: Adzuna API
⚡ Project Highlights
- Built a working AI system combining LLMs, APIs, and backend pipelines
- Handles job matching across multiple listings using scoring logic
- Uses caching and structured processing to improve performance
- Designed modular architecture for scalability
📬 Contact
LinkedIn: LinkedIn Profile Email: satyakeerthidara7@gmail.com