Your AI Agent Is Not a Chatbot. It Should Be a Class.
Last Updated on July 27, 2026 by Editorial Team
Author(s): Gowtham Boyina
Originally published on Towards AI.
Why NVIDIA’s new agent framework treats “prompt engineering” as just… software engineering
Every team building AI agents right now runs into the same wall. You start with a simple prompt. Then you add a tool. Then another tool. Then a memory file. Then a workflow graph to control the order of steps. Before long, your “agent” is spread across five different files, written in three different formats, and nobody on the team can explain the whole system in one sitting.
After the introduction, the article explains why NVIDIA’s NOOA approach argues that agent systems shouldn’t be scattered across prompts, schemas, callbacks, and workflow DSLs—and instead should be expressed as a normal Python class. It shows how docstrings, fields, and typed method signatures map to system prompts, state, and contracts, while “…” methods create agentic behavior through a typed loop that alternates between rendering context, calling the model, executing deterministic Python, and updating state. The piece also outlines NOOA’s memory system (with explicit tools for remembering and retrieving information), discusses how models perform on an evaluation suite and where errors tend to come from execution discipline rather than interface comprehension, and reports results across multiple real benchmarks. Finally, it compares NOOA to other frameworks, notes remaining limitations around safety (requiring sandboxing beyond the interface), and places NOOA in a broader perspective: fewer abstractions, more ordinary software engineering practices—so teams can read, test, and maintain agents more reliably.
Read the full blog for free on Medium.
Join thousands of data leaders on the AI newsletter. Join over 80,000 subscribers and keep up to date with the latest developments in AI. From research to projects and ideas. If you are building an AI startup, an AI-related product, or a service, we invite you to consider becoming a sponsor.
Published via Towards AI
Towards AI Academy
We Build Enterprise-Grade AI. We'll Teach You to Master It Too.
15 engineers. 100,000+ students. Towards AI Academy teaches what actually survives production.
Start free — no commitment:
→ 6-Day Agentic AI Engineering Email Guide — one practical lesson per day
→ Agents Architecture Cheatsheet — 3 years of architecture decisions in 6 pages
Our courses:
→ AI Engineering Certification — 90+ lessons from project selection to deployed product. The most comprehensive practical LLM course out there.
→ Agent Engineering Course — Hands on with production agent architectures, memory, routing, and eval frameworks — built from real enterprise engagements.
→ AI for Work — Understand, evaluate, and apply AI for complex work tasks.
Note: Article content contains the views of the contributing authors and not Towards AI.
Recent Posts
