Talent Activity

The talent activity system provides an audit trail of all talent lifecycle events. Activities are logged automatically via domain events and can also be created manually (e.g., recruiter notes).

Architecture

Domain services emit events via @nestjs/event-emitter. The TalentActivityListener captures these events and persists TalentActivity records. A DataLoader prevents N+1 queries when resolving activities on talent objects.

Cross-service events: Skill activities (SKILL_ADDED) are written directly to the TalentActivity table by background-tasks (Python/AsyncPG) since skills are extracted during CV parsing pipelines. The backend listener handles all other 20 event types.

Activity Types

Type	Trigger
TALENT_CREATED	New talent profile created
TALENT_UPDATED	Profile fields modified
TALENT_DELETED	Talent removed
TALENT_VIEWED	Profile viewed by a user
CV_UPLOADED	CV file uploaded
CV_UPDATED	CV replaced or edited
CV_DELETED	CV removed
CV_PARSED	AI parsing completed
ADDED_TO_POOL	Talent added to a pool
REMOVED_FROM_POOL	Talent removed from a pool
POOL_CHANGED	Talent moved between pools
TALENT_SHARED	Shared with user or company
TALENT_SHARE_REVOKED	Share access revoked
CANDIDATE_CREATED	Added as job candidate
CANDIDATE_STATUS_CHANGED	Candidate status updated
CANDIDATE_STAGE_CHANGED	Moved to different pipeline stage
NOTE_ADDED	Recruiter note added
SKILL_ADDED	Skill added to talent (emitted by background-tasks)
SKILL_REMOVED	Skill removed from talent (not yet implemented)
EXTERNAL_SHARE_CREATED	External share link generated
EXTERNAL_SHARE_ACCESSED	External share link viewed
HIGH_MATCH_DETECTED	High job match score detected

GraphQL Queries

List Activities (Company-Scoped)

query {
  talentActivities(
    filter: {
      activityTypes: [CV_UPLOADED, NOTE_ADDED]
      talentId: "talent-uuid"
      performedById: "user-uuid"
    }
    pagination: { skip: 0, take: 20 }
  ) {
    activities {
      id
      activityType
      description
      metadata
      createdAt
      talent { id firstName lastName }
      performedBy { id firstName lastName }
    }
    totalCount
    hasMore
  }
}

Talent Activity History

query {
  talentActivityHistory(
    talentId: "talent-uuid"
    pagination: { skip: 0, take: 50 }
  ) {
    activities { id activityType description createdAt }
    totalCount
    hasMore
  }
}

Activities on Talent Object

query {
  talent(id: "talent-uuid") {
    id
    firstName
    activities(limit: 10) {
      id activityType description createdAt
    }
  }
}

Manual Activity Creation

Recruiters can create manual activity entries (e.g., interview notes, status updates):

mutation {
  createTalentActivity(input: {
    talentId: "talent-uuid"
    activityType: NOTE_ADDED
    description: "Interview completed - strong technical skills"
    metadata: { rating: 4, stage: "technical" }
  }) { id activityType description createdAt }
}

mutation { deleteTalentActivity(id: "activity-uuid") }

Data Model

Each activity stores talentId, companyId, activityType, performedById, description, and optional metadata (JSON). Client context fields ipAddress and userAgent are captured when available.

Indexes: talentId, companyId, activityType, performedById, createdAt DESC, composites on (companyId, talentId), (companyId, activityType), (talentId, createdAt DESC).

info

Activities cascade-delete when the parent talent or company is removed. The performedById field is set to null if the performing user is deleted.