How shubhanshu2006/sensei works
Sensei is an AI interview-practice and candidate-screening platform for candidates, recruiters, and administrators. Candidates apply to jobs or enter practice interviews; recruiters configure openings and review screened applicants, while admins manage users, credits, and the practice catalog. The frontend communicates with an Express API, whose domain services persist platform records and invoke AI capabilities. The README also describes real-time voice interviews and asynchronous resume screening. Some subsystem links below reflect README-level architecture; unsampled implementation details are not asserted.
Main entry points: ScreeningService.ts Resume screening · client.ts API client · LLMClient.ts LLM provider client · AIEngine.ts AI orchestration
Components 23
User portals
- Candidate portal
frontend/app/(platform)/candidate - Recruiter portal
frontend/app/(platform)/recruiter - Admin portal
frontend/app/(platform)/admin - Live interview UI
frontend/app/(platform)/interview/[sessionId]/page.tsx - API client
frontend/lib/api/client.ts
API and domain
- HTTP application
backend/src/app.ts - API routes
backend/src/routes/index.ts - Application handlers
backend/src/controllers/applications.controller.ts - Application domain
backend/src/services/applications.service.ts - Identity and profiles
backend/src/services/auth.service.ts
AI and interviews
- AI orchestration
backend/src/services/ai/AIEngine.ts - Adaptive interview
backend/src/services/ai/InterviewGraph.ts - Interview evaluation
backend/src/services/ai/EvaluationService.ts - LLM provider client
backend/src/services/ai/LLMClient.ts - Resume extraction
backend/src/services/ai/ResumeParser.ts - Resume screening
backend/src/services/ai/ScreeningService.ts
State and integrations
- Background jobs
backend/src/services/queue - Interview socket server
backend/src/services/websocket/InterviewSocketServer.ts - PostgreSQL
backend/src/database/client.ts - Resume object storage
backend/src/services/storage/S3Service.ts
Other components
- Candidate
- Recruiter
- LLM providers
Connections 22
- Candidate → Candidate portal: uses
- Recruiter → Recruiter portal: uses
- Candidate portal → API client: requests platform data
- Recruiter portal → API client: requests platform data
- Admin portal → API client: manages platform
- Live interview UI → Interview socket server: interviews in real time
- API client → HTTP application: calls API
- HTTP application → API routes: mounts
backend/src/app.ts - API routes → Application handlers: routes applications
- Application handlers → Application domain: calls
backend/src/controllers/applications.controller.ts - Application domain → PostgreSQL: reads and writes
backend/src/services/applications.service.ts - Identity and profiles → PostgreSQL: reads and writes
backend/src/services/auth.service.ts - Application domain → Resume screening: screens applicants
- Resume screening → Resume extraction: extracts resumes
- Resume screening → Background jobs: runs asynchronously
- Resume screening → Resume object storage: uses resume files
- AI orchestration → Adaptive interview: runs interview
backend/src/services/ai/AIEngine.ts - AI orchestration → Resume screening: screens candidate
backend/src/services/ai/AIEngine.ts - Adaptive interview → LLM provider client: generates interview content
- Interview evaluation → LLM provider client: evaluates with models
backend/src/services/ai/EvaluationService.ts - AI orchestration → LLM provider client: transcribes and generates
backend/src/services/ai/AIEngine.ts - LLM provider client → LLM providers: invokes models
backend/src/services/ai/LLMClient.ts