Materials (slides + hands-on tutorials) for Tutorial Day of the Seventh National Research Platform Workshop. The workshop runs May 5–7, 2026 in La Jolla, California; this repository covers Tuesday's full tutorial track plus links to the parallel sessions whose materials are hosted elsewhere. The full 3-day event agenda is on the workshop site.
⏳ The training JupyterHub at training.nrp-nautilus.io is provisioned specifically for 7NRP and will be spun down after the workshop concludes on Thursday, May 7, 2026. Materials in this repository remain available indefinitely. If you'd like to reproduce any tutorial after the workshop — for a class, a follow-up workshop, or your own research — please reach out via Matrix and we'll help you spin up an equivalent environment on persistent NRP infrastructure.
Scan the QR on the right (or open it from https://ucsantacruz.co1.qualtrics.com/jfe/form/SV_3wQP0UrsPXy3nMO?Q_CHL=qr) to take the pre-training survey. It's a quick set of questions about your prior Kubernetes / NRP / AI experience and what you hope to get out of the workshop.
Your answers let us measure how much each tutorial actually moves the needle — comparing pre-training and post-training responses, plus aggregated session telemetry, is how we study the efficacy of our training methods and decide what to keep, cut, or rework for future NRP workshops. The data is collected in aggregate; the more responses we get, the better the next cohort's experience will be.
All Tuesday sessions run at the San Diego Supercomputer Center Auditorium (afternoon parallel-track B in Room 279). Materials hosted in this repository link to a folder; sessions presented by partner organizations link to their own repos.
| Time | Track A — Main auditorium | Track B — Room 279 |
|---|---|---|
| 9:00 – 10:30 AM | 1. NRP and Kubernetes for Education and Research — Tatineni, Sada, Diaz (UCSD/SDSC) | — |
| 10:30 – 11:00 AM | Break | |
| 11:00 AM – 12:00 PM | 2. National Data Platform Education Hub on NRP — Ramonetti, O'Laughlin (UCSD/SDSC) | — |
| 12:00 – 1:00 PM | Lunch | |
| 1:00 – 2:00 PM | 3. Using AI and LLM Inference on NRP — Sada (UCSD/SDSC) | Tutorial on Drone Usage and IoT — Graham (UCSD/SDSC) |
| 2:00 – 3:00 PM | 4. Setting Up Custom JupyterHubs — Sada, Diaz (UCSD/SDSC) | ICICLE Digital Agriculture Pipeline — Subramoni (Ohio State) |
| 3:00 – 3:30 PM | Break | |
| 3:30 – 5:00 PM | 5. DoE American Science Cloud Experiments with NRP — Balcas, Yang (ESnet) | AgTech for the Curriculum Discussion — Wuerthwein (UCSD/SDSC) |
Track A sessions all run in the same JupyterHub at training.nrp-nautilus.io — sign in once and you're set up for every Track A tutorial. Materials for Track B sessions are maintained by their respective presenters and are not in this repository; contact the presenters directly.
Two ways to follow along during the workshop:
Option 1 — Training JupyterHub (recommended, zero install). The workshop hub at training.nrp-nautilus.io is pre-configured: every spawned JupyterLab pod already has kubectl installed and a kubeconfig wired up to the same identity, so you can open a terminal and run kubectl immediately. Click below to clone this repo straight into your JupyterLab session:
Option 2 — kubectl on your laptop. Install kubectl (Linux / macOS / Windows) and use the ready kubeconfig at files/nrp-training.kubeconfig. It carries the jupyterhub-sa service-account token, cluster CA, and nrp-training-k8s as the default namespace; the embedded token is valid for the duration of 7NRP, through end-of-day Thursday, May 7, 2026. Step-by-step instructions live in Tutorial 1 → Interacting with NRP.
Conventions
- Tutorials 1 and 3 (AI/LLM) share the
nrp-training-k8snamespace. It already exists for the workshop; if you need to recreate it later:kubectl create namespace nrp-training-k8s. - Tutorial 4 (Custom JupyterHubs) uses pre-created per-participant namespaces (
nrp-training-000…nrp-training-099). - Replace
<username>in any YAML or command with your NRP or GitHub username to avoid name collisions. - Every workload must declare CPU and memory
requestsandlimits— a cluster-wide Gatekeeper policy rejects pods without them.
Click anywhere on a card (number, title, path, or technology badge) to open the materials.
1_nrp_kubernetes_education_research/
Part 1 — intro slides: NRP-Overview-Jupyter.pdf (NRP overview + JupyterHub orientation, presented before the hands-on).
Part 1 — hands-on: the markdown tutorial walks through every core Kubernetes object researchers and educators need: pods, PersistentVolumeClaims, multi-container (sidecar) pods, ConfigMaps + Secrets, Deployments, Batch Jobs, kubectl cp / port-forward / patch, HTTPS-fronted Services via Ingress, scheduling primitives (taints, tolerations, labels, affinity) for the workshop's reserved GPU pool, and finally a GPU pod.
Pedro Ramonetti and Katie O'Laughlin walk through the National Data Platform's Education Hub running on NRP — what it offers, how courses and workshops use it, and the integration patterns that wire NDP datasets into JupyterHub-based curricula. Slides only; the live session walks through running NDP demos on the same workshop JupyterHub.
Use NRP's GPU-backed managed LLM service: hit the OpenAI-compatible endpoint from Jupyter AI, curl, and Python (openai SDK); drive an agentic coding TUI (opencode) against it to generate a working chess game; then bring up your own GPU pod for a PyTorch sanity check, a TGI inference server, and a Milvus RAG pipeline answering questions over the NRP documentation.
3_custom_jupyterhubs_classroom_research/
Deploy your own JupyterHub for a course or research group: install Helm, add the JupyterHub Helm chart, configure Ingress, set up multiple image profiles and per-profile resource limits, attach shared storage, and learn how to build custom container images via NRP GitLab CI/CD.
Justas Balcas and Xi Yang from ESnet present the DoE American Science Cloud vision and its overlap with NRP, then walk through hands-on examples against the DOE Integrated Research Infrastructure (IRI). Slides are in this repo; the working code lives in doe-iri/iri-facility-api-examples — clone it from a JupyterLab terminal in the same training JupyterHub used by the rest of Tutorial Day.
Three Tuesday-afternoon Track B sessions are presented by partner organizations and their materials live in their own repositories or are distributed by the presenters.
| Time | Session | Speaker | Where to find materials |
|---|---|---|---|
| 1:00 – 2:00 PM | Tutorial on Drone Usage and IoT | John Graham (UCSD/SDSC) | Contact the speaker / Matrix |
| 2:00 – 3:00 PM | Tutorial Using ICICLE Digital Agriculture Pipeline | Hari Subramoni (Ohio State) | icicle-ai on GitHub |
| 3:30 – 5:00 PM | AgTech for the Curriculum Discussion | Frank Wuerthwein (UCSD/SDSC) | Discussion session — no slides |
| 📚 User documentation | Getting started on NRP |
| 💬 Matrix chat | Live support during the workshop, and for post-workshop reproduction requests |
| 🗓️ 7NRP event site | Schedule, venue, registration |
| 📋 Full 3-day agenda | All sessions Tue–Thu (Tutorial Day, plenary talks, panels) |
| 🚀 Tutorial JupyterHub | Workshop JupyterHub instance — spun down after May 7, 2026 |