research-computing

Open OnDemand on Eureka — Apache + OIDC + Let's Encrypt, pre-loaded with Jupyter, RStudio, VS Code, MATLAB, and ParaView. Reproducible deployment for the AMII × UAlberta PAICE cluster.

Helm chart for self-hosted RAGFlow — chat with your documents using AI, backed by vector search, OIDC auth, and a full-stack data pipeline. Drop it on Kubernetes and go.

HTCondor research compute platform on Proxmox VE. Researcher self-service portal with quota enforcement, Apptainer job sandboxing, NFS per-project data isolation, SLO-driven observability (Prometheus + Grafana), and full IaC via Terraform + Ansible. Zero manual deployment.

Kubernetes DaemonSet deployment of Squid for the Vulcan HPC cluster at the University of Alberta, forming a peer-coordinated cache for CVMFS and research software distribution with optimized caching, peering, and domain whitelisting.

Open WebUI as an Open OnDemand Batch Connect app: a full LLM chat interface running in the user's own batch job, from a self-contained container. Endpoint-agnostic, with no model, endpoint or API key built in, and per-user chat history that persists between sessions.

Amarel VS Code Skill: set up VS Code Remote-SSH on Rutgers Amarel and fix the GLIBC 2.28 error on macOS or Windows.

Open repository of test scripts for future procurements.

Research-oriented analysis of BRCA1-related genomic data using statistical and reproducible computational workflows (ITCGA Summer Workshop project).