Mentors
How we staff and run rooms
We hire for traceable shipping experience, not keynote counts. Each section below maps to artifacts mentors must maintain.
Mentor standards
- Two years minimum operating production ML systems with documented incidents.
- Public sample of runbooks or model cards redacted for confidentiality.
- Willingness to teach on-camera from Gangwon labs at least once per quarter.
Learning workflow
Async prep → paired implementation → critique → field rehearsal → handoff pack. No stage is skipped even when sponsors request shortcuts.
Classroom tools
Queue board
Transparent GPU tokens and cold-start warnings updated every lab block.
Rubric printer
Physical handouts for labeling sessions to reduce screen fatigue.
Incident hotseat
Dedicated desk for paging mentors during deployment exercises.