CareerPrep AI is an AI-powered career preparation platform that generates personalized career preparation guides for different technical roles.
The system combines Large Language Models (LLMs), Retrieval Augmented Generation (RAG), LangGraph workflow orchestration, and vector databases to create structured career preparation material.
Users can generate complete preparation guides containing:
- Role overview
- Required skills
- Learning roadmap
- Interview preparation
- Project recommendations
- Learning resources
- Downloadable PDF guides
Users provide a target role such as:
- AI Engineer
- Software Engineer
- Data Engineer
- DevOps Engineer
- Full Stack Developer
The system generates a complete preparation guide containing:
Includes:
- Role explanation
- Responsibilities
- Industry expectations
- Required technical knowledge
Includes:
- Beginner concepts
- Intermediate skills
- Advanced topics
- Practical learning milestones
Includes:
- Technical questions
- Conceptual questions
- Role-specific interview preparation
- System design topics
Includes:
- Beginner projects
- Intermediate projects
- Advanced portfolio projects
Includes:
- Documentation
- Courses
- Books
- Practice resources
LLMs are used for:
- Understanding user requirements
- Generating career guides
- Answer generation
- Content structuring
Supported providers:
- Groq LLM
- HuggingFace Models
LangChain is used for:
- LLM integration
- Prompt management
- Retrieval pipelines
- Document processing
- Output formatting
Main components:
- Prompt Templates
- LLM Chains
- Retrievers
- Document loaders
- Output parsers
CareerPrep AI uses LangGraph for workflow orchestration.
The system does not use independent AI agents. Instead, it uses multiple specialized RAG processing nodes running inside a LangGraph workflow.
CareerPrep AI uses multiple Retrieval Augmented Generation pipelines.
Each RAG module focuses on a specific part of the career guide.
Generates:
- Learning paths
- Required technologies
- Skill progression
Generates:
- Technical interview questions
- Concept explanations
- Interview preparation material
Generates:
- Portfolio projects
- Implementation ideas
- Difficulty levels
Generates:
- Documentation
- Courses
- Books
- Learning resources
RAG improves generation quality by providing relevant external knowledge to the LLM.
Workflow:
User Query
↓
Query Processing
↓
Vector Similarity Search
↓
Relevant Document Retrieval
↓
Context Injection
↓
LLM Generation
↓
Final Response
Benefits:
- Reduces hallucination
- Provides domain-specific knowledge
- Improves response accuracy
- Uses custom knowledge sources
CareerPrep AI uses DataStax AstraDB as the vector database for storing and retrieving domain-specific knowledge used by the RAG pipelines.
The vector database stores document embeddings along with custom metadata to enable more accurate and targeted retrieval.
Used for:
- Storing document embeddings
- Semantic similarity search
- Context retrieval for RAG pipelines
- Metadata-based filtering
- Knowledge separation across different career roles and guide sections
Architecture:
Documents
↓
Text Chunking
↓
Embedding Generation
↓
AstraDB Vector Store
↓
Similarity Search
↓
LLM Context
Used for:
- REST API development
- Request handling
- Backend services
- API routing
Backend structure:
backend/
│
├── app/
│
├── api/
│ ├── guide.py
│ └── health.py
│
├── nodes/
│ ├── roadmap_RAG.py
│ ├── interview_RAG.py
│ ├── projects_RAG.py
│ └── resources_RAG.py
│
├── retrievers/
│
├── prompts/
│
├── graph.py
│
└── main.py
Used for:
- UI components
- User interaction
- State management
Used for:
- Type safety
- Maintainable frontend development
Used for:
- Fast development
- Production builds
Used for:
- Responsive UI
- Modern styling
Additional libraries:
- Axios → API communication
- Framer Motion → Animations
- React Markdown → Markdown rendering
CareerPrep AI generates downloadable PDF preparation guides.
Features:
- Structured formatting
- Markdown conversion
- Professional document layout
Example:
AI_Engineer_Career_Guide.pdf
Software_Engineer_Career_Guide.pdf
Example:
# AstraDB
ASTRA_DB_API_ENDPOINT=
ASTRA_DB_APPLICATION_TOKEN=
# LLM
GROQ_API_KEY=
HF_TOKEN=
Never commit environment variables or secrets to GitHub.
git clone https://github.com/Noore-hira/CareerPrep_AI.git
cd CareerPrep_AIcd backendCreate environment:
python -m venv .venvInstall dependencies:
pip install -r requirements.txtRun backend:
uvicorn app.main:app --reloadBackend:
http://localhost:8000
Swagger:
http://localhost:8000/docs
cd frontendInstall dependencies:
pnpm installRun:
pnpm devFrontend:
http://localhost:5173
Frontend is deployed on:
Vercel
Deployment process:
GitHub Repository
↓
Vercel Build
↓
React Production Bundle
↓
Hosted Frontend
Production build:
pnpm buildBackend is deployed on:
Red Hat OpenShift Developer Sandbox
Deployment architecture:
FastAPI Application
↓
Docker Container
↓
OpenShift Deployment
↓
Running Backend Pod
↓
Public API Route
Production command:
uvicorn app.main:app --host 0.0.0.0 --port $PORTEnvironment variables are configured through OpenShift.
Implemented:
- Environment variable management
- API key handling
- CORS configuration
Production CORS should restrict allowed origins:
allow_origins=[
"https://your-vercel-domain.com"
]Possible enhancements:
- User authentication
- Saved career plans
- Progress tracking
- AI mock interviews
- Resume analysis
- Personalized learning paths
- Voice interview simulation
Developed as an AI Engineering project demonstrating:
- Generative AI applications
- Retrieval Augmented Generation
- LangGraph workflows
- LLM applications
- Full-stack AI development
| Category | Technology |
|---|---|
| Frontend | React, TypeScript, Vite, Tailwind CSS |
| Backend | FastAPI |
| AI Framework | LangChain |
| Workflow Orchestration | LangGraph |
| Retrieval System | Multi-RAG Architecture |
| Vector Database | AstraDB |
| LLM Providers | Groq, HuggingFace |
| API Communication | Axios |
| PDF Generation | ReportLab |
| Frontend Deployment | Vercel |
| Backend Deployment | Red Hat OpenShift |