How pavankl24/ai_interview_agent works
COGNIV AI is a technical interview practice platform. Candidates configure an interview, answer adaptive AI-generated questions, and receive a final report saved for later review; they can also build a personalized learning roadmap from interview results. The graph includes the React client, FastAPI interview and learning flows, AI/reporting, and persistence. README capabilities such as Judge0 code execution and Supabase authentication are included; their detailed wiring is not fully shown in sampled source, so those links are documented or marked optional.
Main entry points: App.tsx Interview workspace · api.py Interview orchestration · database.py Interview persistence · learning_service.py Roadmap generation
Components 21
Candidate experience
- Interview workspace
web/src/App.tsx - HTTP client
web/src/apiClient.ts - Speech input
web/src/hooks/useSpeechRecognition.ts - Question timer
web/src/hooks/useCountdown.ts - Copy utility
web/src/hooks/useCopyToClipboard.ts - Roadmap interface
web/src/LearningRoadmap.tsx
Interview services
- API assembly
backend/main.py - Interview orchestration
api.py - Interview routes
backend/routers/interview.py - Agent facade
agent.py - Adaptive interviewer
backend/services/ai_service.py - Performance report
report.py
Learning roadmap
- Learning routes
backend/routers/learning.py - Roadmap generation
backend/services/learning_service.py
Identity and persistence
- Supabase client
web/src/supabase.ts - Session storage
backend/services/session_store.py - Interview persistence
database.py
Other components
- Candidate
- Groq LLM
- Supabase services
- Judge0 executor
Connections 29
- Candidate → Interview workspace: configures and answers
Readme.md - Interview workspace → HTTP client: sends requests
web/src/App.tsx - Interview workspace → Speech input: uses speech input
web/src/App.tsx - Interview workspace → Question timer: uses countdown
web/src/App.tsx - Interview workspace → Copy utility: uses copy
web/src/App.tsx - Interview workspace → Roadmap interface: renders roadmap
web/src/App.tsx - Interview workspace → Supabase client: uses authentication
web/src/App.tsx - HTTP client → API assembly: calls API
Readme.md - API assembly → Interview routes: mounts routes
backend/main.py - API assembly → Learning routes: mounts routes
backend/main.py - Interview routes → Adaptive interviewer: uses interviewer
backend/routers/interview.py - Interview orchestration → Agent facade: creates agent
api.py - Agent facade → Adaptive interviewer: re-exports
agent.py - Interview orchestration → Performance report: generates report
api.py - Interview orchestration → Session storage: loads and saves sessions
api.py - Interview orchestration → Interview persistence: reads history
api.py - Adaptive interviewer → Groq LLM: requests completions
Readme.md - Performance report → Groq LLM: generates evaluation
report.py - Performance report → Interview persistence: report persistence
- Learning routes → Roadmap generation: dispatches requests
backend/routers/learning.py - Learning routes → Interview persistence: verifies identity
backend/routers/learning.py - Roadmap generation → Interview persistence: reads interview reports
backend/services/learning_service.py - Roadmap generation → Groq LLM: generates learning content
Readme.md - Roadmap interface → HTTP client: requests roadmap
- Interview persistence → Supabase services: stores sessions
Readme.md - Supabase client → Supabase services: authenticates users
Readme.md - Interview workspace → Interview orchestration: submits code
Readme.md - Interview orchestration → Judge0 executor: dispatches execution
Readme.md - Roadmap generation → Judge0 executor: validates exercises