Retour du Flutter Paris avec une spéciale IA

C'est le grand retour du Flutter Paris 🥳 ! Après quelques mois en sommeil, nous profitons de cette rentrée pour lancer une nouvelle saison. Pour l'occasion, nous serons accueilli chez Ippon Technologies pour découvrir deux présentations 1️⃣ 1ère présentation 🇬🇧 (Sasha Denisov) Edge AI in Flutter: How Far Have We Come? When fluttergemma shipped its first release two years ago, on-device AI in Flutter was real but narrow: a MobileNet-class classifier for vision, and a small, text-only LLM wrapped over MediaPipe that answered one prompt at a time — no images, no tools, no memory. In 2026 the picture is almost unrecognizable: your phone runs a multimodal open model — Gemma, Qwen, Llama, and more — that sees images, calls tools, remembers context, and retrieves from a local knowledge base, entirely offline, with nothing ever leaving the device. This talk is a retrospective on how we got here — across the models, the runtimes that execute them (Google's LiteRT-LM and MediaPipe, other engines like llama.cpp and ONNX Runtime, and even the model the OS now ships itself), and the Dart packages that made all of it native to Flutter. As a throughline, we'll follow fluttergemma: how it grew, release by release, from a thin MediaPipe wrapper around a 2B model into a modular ecosystem of its own — pluggable engines, embeddings, on-device RAG, and agents that call tools, run skills, and even speak MCP — and where it goes next. Along the way we'll trace what changed in the silicon (CPU → GPU → NPU) and the developer experience, and look honestly at what's production-ready today. 2️⃣ 2ème présentation 🇬🇧 (Elaine Dias Batista) Flutter development in the AI era Description communiquée ultérieurement 🎁 Surprise, surprise Nous aurons également une très bonne nouvelle à vous partager…