Large language models can now power capable software agents, yet real‑world success comes from disciplined engineering rather than flashy frameworks. Most reliable agents are built from simple, composable patterns instead of heavy abstractions.

The talk will introduce several patterns that add complexity / autonomy only when it pays off:

1. Augmented LLM (retrieval, tools, memory) as the atomic building block
2. Workflow motifs: prompt chaining, routing, parallelization etc with concrete criteria and implementation tips
3. Autonomous agents that loop through plan‑act‑observe‑reflect cycles to tackle open‑ended tasks

Attendees will leave with a practical decision framework for escalating from a single prompt to a multi‑step agent, reference implementations they can reproduce in a few lines of code, and robust guardrails for shipping trustworthy, cost‑effective agents at scale.