AI Agents in Python

Ever wonder what people really mean by “AI agents,” and how they’re different from a simple LLM call? In this talk we’ll demystify agents, look at the specs and patterns behind them, and then get hands-on with Python—consuming agents locally and over the network, building your own in-process agent, and hosting one for remote use.

What we’ll cover

• What an agent is (and isn’t): core concepts, capabilities, limitations, and when to use one.
• Relevant specs & patterns: tool/function calling, planning/act/observe loops, state & memory, safety/guardrails, and common protocol shapes you’ll encounter.
• Building a local agent: clean tool interfaces, pydantic-style I/O schemas, orchestration loops, and testing.
• Hosting a remote agent: FastAPI-style endpoints, OpenAPI/JSON schema contracts, streaming, auth/rate-limits, logging & observability.

Who should attend
Python devs, ML/AI engineers, and builders who want practical patterns for using agents in real applications—no prior agent framework experience required.

Takeaways
You’ll leave with a clear mental model of AI agents, a reference architecture for local vs. remote setups, and copy-pasteable patterns to start integrating agents into your Python projects.

#VegasTech