How ship1220/intervista_updated works
This is an AI job-preparation application for interview practice and personalized course generation. Its user-facing workflow is served by the FastAPI application: users provide interview or career context, practice and submit answers, and receive feedback; the application also supports audio transcription and course/resource features. Internally, interview chains can retrieve supporting material from a RAG index, while a separate resource-discovery pipeline and contextual bandit service support learning and adaptation. The README only names the product, and sampled excerpts are partial, so uncertain connections between API routes and services are omitted.
Main entry points: rag_pipeline.py RAG coordinator · pipeline.py Resource discovery · main.py FastAPI application · query_generator.py Search query generation
Components 22
Application
- FastAPI application
main.py - Audio transcription
speech/transcription.py
Interview Practice
- Interview chains
core/chains/base_chain.py - Prompt templates
core/prompts/prompt_manager.py - Answer evaluation
evaluation/evaluator.py - Skill profile
user_skill_profile.py - Skill integration
user_skill_integration.py - Adaptive selection
services/rl/rl_service.py
Learning Support
- RAG coordinator
services/rag/rag_pipeline.py - Document ingestion
services/rag/document_ingester.py - Document retrieval
services/rag/retriever.py - Vector index
services/rag/vector_store.py - RAG seed data
services/rag/seed_data.py - Resource discovery
services/learning_resources/pipeline.py - Search query generation
services/learning_resources/query_generator.py - Resource fetchers
services/learning_resources/fetchers - Resource normalization
services/learning_resources/normalizer.py - Resource ranking
services/learning_resources/ranker.py
Shared Services
- LLM service
core/llm/llm_service.py - Application database
database.py
Other components
- Candidate
- Groq API
Connections 18
- Candidate → FastAPI application: uses
- FastAPI application → Application database: reads and writes
- FastAPI application → Audio transcription: transcribes audio
- Interview chains → Document retrieval: retrieves context
- Interview chains → Prompt templates: formats prompts
- LLM service → Groq API: requests completions
- Search query generation → LLM service: generates queries
- Search query generation → Prompt templates: uses templates
- RAG coordinator → Document ingestion: processes documents
- RAG coordinator → Vector index: indexes documents
- RAG coordinator → Document retrieval: coordinates retrieval
- RAG coordinator → RAG seed data: loads seed data
- Document retrieval → Vector index: searches index
- Document ingestion → Vector index: supplies chunks
- Resource discovery → Search query generation: generates searches
- Resource discovery → Resource fetchers: retrieves resources
- Resource discovery → Resource normalization: normalizes results
- Resource discovery → Resource ranking: ranks results