7 agents in this category
Analyses interview recordings for authenticity (reading/recited vs spontaneous). Advisory delivery signal — not a hiring verdict or proof of misconduct.
Labels each turn of a pre-diarized date-conversation transcript with the v3 bid/response dialogue-act dimensions (task_function, connection, close, direction, build), writing them to a cubby for the autoresearch eval loop.
On-platform hiring agent: role-rubric resume scoring, 9-lens interview transcript analysis, voice analysis, and a compounding learning loop. All inference via internal models. A/B example: two engagements (control vs experimental resume prompt) split 50/50 by weight; per-Job assignment, not sticky.
On-platform hiring agent: role-rubric resume scoring, 9-lens interview transcript analysis, voice analysis, and a compounding learning loop. All inference via internal models.
Layer 1 base-emotion labelling for conversations (nlp-kit emotion.basic, on-platform models).
GitHub engineering intelligence agent that writes activity into a vault cubby and optional Notion execution log.