IA Générative

LLM Systems: Architecture First, Framework Second

LangChain, LangGraph, Claude ADK, Google ADK, PydanticAI — different names, different abstractions, same promise. This article cuts through the framework noise and identifies the five architectural properties that actually determine whether an LLM system survives production.

Par

Cyril Noirot

1 avril 2026

6 min de lecture

If you have spent even a few hours in the GenAI ecosystem recently, you have probably hit the same wall.

Every framework you encounter is just a different answer to that question. The sooner you internalize this, the sooner the landscape stops being overwhelming and starts being legible.

1. Model layer OpenAI, Claude, Gemini. This is a function: input → output. Nothing more.

2. Orchestration layer How calls are chained, routed, retried. This is where LangChain, LangGraph, and the ADKs sit.

À propos de l'auteur

Cyril Noirot

Lead Data Scientist

Data scientist freelance. Je conçois et déploie des systèmes de décision — prévision, pricing, marketing measurement, optimisation.

Plus sur Cyril Réalisations LinkedIn

From SaaS to intelligence native: the feedback loop.

Intelligence-native systems need agent access to decision artefacts and feedback loops. Why context, not models, is the differentiator — and how MCP, traditional ML, and versioned artefacts fit together.

IA Générative8 min

Who controls what your AI recommends?

Seasons, campaigns, and weekly events — retail runs on overlapping cycles, and the AI recommender has to keep up with all of them. Notes on the business-rules control surface that lets merchandising teams steer a conversational recommender without editing prompts, filing tickets, or waiting for a deploy.

IA Générative11 min

Ce que le commerce agentique exige vraiment en production

Macy's annonce que les clients utilisant son assistant IA dépensent 4,75x plus. Sephora vient de lancer une app dans ChatGPT. Zalando déploie son assistant dans 25 marchés. La question pour tous les autres retailers n'est plus 'faut-il le faire ?' mais 'comment l'architecturer pour que ça tienne en production ?'

Newsletter

Articles techniques sur la prévision, le pricing et les systèmes de décision. Aucune fréquence imposée.

Enter your email