distill-my-mentor

To create an AI mentor from academic chat logs and annotated papers, you run a multi-phase guided workflow that ingests your mentor interactions, extracts their style and guidance context, and assembles a runnable conversational Skill on disk for immediate use.

Mentor distillation in academic research advising is the process of transforming real interactions like emails, chat logs, and paper track changes into a conversational AI that can advise, review papers, and rewrite text in your mentor's voice.

Yes, you can use PDF literature and email threads to emulate your mentor's academic writing style. The Skill supports multi-source data import from chat logs, paper annotations, and literature PDFs to improve the fidelity of the generated mentor persona.

You can correct the behavior of a generated mentor AI during a conversation by providing immediate in-conversation style corrections, such as indicating "they wouldn't say that," which updates the AI's behavior instantly without restarting the session.

When distilling a mentor Skill from descriptive-only intake, the primary limitation is reduced fidelity compared to ingesting actual chat logs or paper annotations, as the generated AI relies entirely on your subjective descriptions rather than concrete mentor interaction data.

Yes, the mentor distillation workflow supports paper review with track changes. It provides an operational mode specifically for annotation-style paper review, allowing the distilled mentor AI to review literature and provide feedback mimicking your mentor's approach.