Now onboarding early-access partners
Data science workspaces that are ready the moment class starts.
KernelQuad gives university courses and analyst teams browser-based Python, R and SQL environments with shared GPUs, hard spending caps and data that stays in the region you choose.
Launching in US East, EU (Frankfurt) and Asia Pacific (Mumbai).
Course
Week 6 · CNNs
Datasets
Submissions
Environment
Workspace
GPU · ¼ L4
Python 3.12
PyTorch
import torch
from torchvision.models import resnet18
model = resnet18(num_classes=10).cuda()
# course dataset, mounted read-only
data = load("/datasets/ml-401/cifar10")
epoch 3/10 · loss 0.842 · acc 71.4%
GPU slice in use¼ NVIDIA L4
Your term budget$4.10 of $15.00
Works with
JupyterLabVS CodePythonRSQLPyTorchTensorFlowSparkNVIDIA GPUs
Why KernelQuad
Free GPU programmes for students are shrinking, and cloud bills are unpredictable. KernelQuad is built around how courses and teams actually work.
Set up once, for the whole class
An instructor picks a template — Python, R, PyTorch, Spark — and every student gets the same environment. No installs, no "works on my laptop".
GPUs without surprise bills
Fractional GPUs let several students share one card for coursework. Per-student and per-course budgets are hard caps, and idle machines shut down automatically.
Campus hardware plus cloud
Run everyday workloads on GPU workstations in your own lab and burst to the cloud when deadlines spike — one platform, one bill.
Your data stays where you put it
Choose a region when you create a workspace. Your notebooks and data are stored in that region. How residency works →
How it works
From a course outline to a running GPU notebook for every student.
Choose a template
Start from a curated environment or bring your own container image and package list.
Set budgets and region
Pick where data lives and the most each student or team member can spend. Caps are enforced, not advisory.
Invite users
Students sign in from your LMS or SSO; analysts join by invitation. Everyone lands in a ready workspace.
Work, share and grade
Run notebooks on CPU or GPU, share results, collect submissions and re-run any work exactly as it ran.
Built for two audiences
Universities
Courses, labs and research groups
- Sign in from Canvas, Moodle or your campus SSO
- Assignments, shared datasets and reproducible environments
- Cost per student per term, agreed up front
Analysts
Analyst and data teams
- Notebooks, SQL and Python in one workspace
- Connect warehouses, databases and files
- Region-pinned data for regulated work
Graduates
Keep your work after you graduate
Students keep their workspace and portfolio on a low-cost personal plan, and can bring it with them to their first analyst role.
Launch regions
Workspaces can be created in three regions at launch. More will follow based on demand.
us-east-1
United States
US East (N. Virginia)
eu-central-1
European Union
EU (Frankfurt) — GDPR-aligned data residency
ap-south-1
India
Asia Pacific (Mumbai)
Frequently asked questions
Can a course really never go over budget?
Yes. Budgets are hard caps. When a student or course reaches its limit, GPU sessions pause and the workspace stays available on CPU so no work is lost. Instructors can raise a cap at any time.
What does a "fractional GPU" mean?
Most coursework does not need a whole data-centre GPU. We partition a GPU so that several students share it with isolated memory, which makes GPU time far cheaper per student. Full GPUs are available for projects and research.
Do we need to buy hardware?
No. KernelQuad runs fully in the cloud. Departments that want predictable costs can add on-campus GPU workstations that we manage, with cloud capacity used only for peaks.
Where is our data stored?
In the region you choose when you create a workspace. Encrypted backups are kept with a second provider. See security & data residency for details.
How do you prevent misuse of GPUs?
Every account is tied to a verified institution or paying customer, usage is capped by budget, and we monitor for prohibited workloads such as cryptocurrency mining. See our acceptable use policy.
Can students keep their work after the course ends?
Yes. At the end of a term students can export their work or move it to a personal plan.
Become an early-access partner
We're working with a small number of instructors, departments and analyst teams ahead of general availability.