How mubashranoor04/ai-powered-hr-recruitment- works
This is a FastAPI-backed HR recruitment RAG platform with a Streamlit frontend. Recruiters authenticate, organize documents in rooms, upload files for extraction and embedding, then ask room-scoped questions and receive grounded answers with source citations; chat history is retained. The graph separates API orchestration, identity and workspace management, ingestion, retrieval/generation, and persistence. Source excerpts establish most backend calls; the frontend connection and unsampled formats are included only where README documentation supports them.
Main entry points: database.py SQLAlchemy sessions · main.py FastAPI app · upload.py Upload endpoint · auth.py JWT and passwords
Components 16
API orchestration
- Streamlit frontend
streamlit_app.py - FastAPI app
app/main.py
Identity and rooms
- Auth endpoints
app/routers/auth.py - JWT and passwords
app/services/auth.py - Room endpoints
app/routers/rooms.py
Document ingestion
- Upload endpoint
app/routers/upload.py - Format extractors
app/services/ingestion - Chunk and embed
app/services/embedder.py
Retrieval and answers
- Qdrant vectors
- Chat endpoints
app/routers/chat.py - RAG service
app/services/rag.py - Groq LLM
Persistence
- SQLAlchemy sessions
app/db/database.py - Domain records
app/db/models.py - Relational database
Other components
- Recruiter
Connections 24
- Recruiter → Streamlit frontend: uses
README.md - Streamlit frontend → FastAPI app: calls API
README.md - FastAPI app → Auth endpoints: registers routes
app/main.py - FastAPI app → Room endpoints: registers routes
app/main.py - FastAPI app → Upload endpoint: registers routes
app/main.py - FastAPI app → Chat endpoints: registers routes
app/main.py - Auth endpoints → JWT and passwords: hashes and signs
app/routers/auth.py - Room endpoints → JWT and passwords: checks identity
app/routers/rooms.py - Upload endpoint → JWT and passwords: checks identity
app/routers/upload.py - Chat endpoints → JWT and passwords: checks identity
app/routers/chat.py - Auth endpoints → SQLAlchemy sessions: queries session
app/routers/auth.py - Room endpoints → SQLAlchemy sessions: queries session
app/routers/rooms.py - Upload endpoint → SQLAlchemy sessions: records files
app/routers/upload.py - Chat endpoints → SQLAlchemy sessions: stores history
app/routers/chat.py - JWT and passwords → SQLAlchemy sessions: looks up users
app/services/auth.py - SQLAlchemy sessions → Domain records: maps records
app/db/models.py - SQLAlchemy sessions → Relational database: connects
README.md - Upload endpoint → Format extractors: dispatches formats
app/routers/upload.py - Upload endpoint → Chunk and embed: chunks and stores
app/routers/upload.py - Chunk and embed → Qdrant vectors: upserts vectors
app/services/embedder.py - Chat endpoints → RAG service: requests answer
app/routers/chat.py - RAG service → Chunk and embed: encodes query
app/services/rag.py - RAG service → Qdrant vectors: retrieves chunks
app/services/rag.py - RAG service → Groq LLM: generates response
app/services/rag.py