For universities
Give every student a GPU-ready workspace — on a budget you set.
KernelQuad is built around how courses actually run: a fixed term, a known number of students, and deadlines that all land on the same week.
How a course works
Create the course
Pick a template, attach datasets and set a budget per student for the term.
Students sign in
From your LMS or campus SSO. Everyone gets an identical, persistent environment.
Teach and grade
Distribute notebooks, collect submissions and re-run any student's work exactly as they ran it.
Stay on budget
Hard caps and idle shutdown mean the course never costs more than agreed.
What's included
Teaching tools
- Course templates for Python, R, PyTorch, TensorFlow and Spark
- Assignment distribution and collection
- Shared, read-only course datasets
- Instructor and teaching-assistant roles
Compute that fits coursework
- Fractional GPUs: several students share one card
- Full GPUs available for projects and theses
- Automatic shutdown of idle sessions
- Per-student, per-course and per-department caps
Campus integration
- LMS integration (LTI 1.3) for Canvas, Moodle and others
- SAML / OIDC single sign-on
- Optional on-campus GPU workstations managed by KernelQuad
- Cloud burst during peak weeks
KernelQuad is in early access. Some features listed here are in active development; we will confirm exactly what is available for your term during onboarding.