A hands-on tutorial showing how to configure NVIDIA NeMo Agent Toolkit with Docker Model Runner (DMR) for local LLM inference, with a focus on adding observability to AI agents via OpenTelemetry tracing. Covers setting up DMR with the smollm2 model, configuring NeMo via YAML with a ReAct agent and Wikipedia search tool, and wiring up an OTel collector to capture spans from agent runs.
68 Impressions