Hands-on course teaching production LLM serving on Kubernetes — from synthetic data and LoRA fine-tuning through model packaging, three serving patterns (plain vLLM, KServe, vLLM Router), observability, autoscaling, and GitOps.
Companion course (AgentOps): https://github.com/schoolofdevops/303-agentops
| Version | Branch/Tag | Content |
|---|---|---|
| v1.0.0+ (current) | main |
LLMOps-only: Labs 00-06 + new serving patterns. AgentOps removed. |
| v0.19.0 (archived) | v0.19.0 tag |
Combined LLMOps + AgentOps (Labs 00-13) — 3-day workshop |
| v0.19.x (maintenance) | v0.19.x branch |
Bug fixes for v0.19.0 students during the transition window |
| AgentOps | 303-agentops | Hermes Agent, MCP tools, Kubernetes Agent Sandbox — companion course |
If you enrolled in the Udemy course on the combined LLMOps + AgentOps workshop: pin your fork to the v0.19.0 tag or v0.19.x branch. The v0.19.0 content is permanently available there.
If you want the latest LLMOps-only course: you are on the right branch (main). v1.0.0 covers multiple serving patterns (plain vLLM Deployment, KServe InferenceService, vLLM Router), disk-based model loading, and a full Argo Workflows training pipeline.
If you want AgentOps content: https://github.com/schoolofdevops/303-agentops — builds on this course as a prerequisite.
https://llmops.schoolofdevops.com (Docusaurus — built from course-content/)
See Lab 00 in the course site for cluster setup. Requires Docker Desktop + KIND + 16GB RAM.