Decision Models SF | Demo Night

Join Liquid AI and the San Francisco AI community for an evening dedicated to decision models: purpose-built models that make fast, structured decisions without generating tokens.

Unlike language models that produce free-form text, decision models return calibrated probabilities for outcomes such as:

• Noul: Should this action happen?

• Choice: Which option should the system select?

• Score: How should this input be rated?

Not every AI task needs generation. Routing requests, approving agent actions, moderating content, evaluating outputs and inspecting images often require a clear, reliable decision, not another paragraph of text.

Decision models offer a new way to build these systems: predictable structured outputs, calibrated confidence, low latency and zero generated tokens. They can work independently or alongside LLMs as fast decision layers for agents and production applications.

Come learn how they work, explore where they fit in the modern AI stack and see what developers are already building with Liquid's decision model family across text, vision and audio.

Agenda

5:30–6:00 PM — Doors open, food, and networking
6:00–7:00 PM — Everything you need to know about decision modes with Maxime Labonne
7:00–8:00 PM — Decision Model Showcase
8:00–9:00 PM — Networking

Presenters

Maxime Labonne — Head of Post-Training at Liquid AI
Maxime leads post-training at Liquid AI. He is the creator of the open-source LLM Course and co-author of the bestselling LLM Engineer's Handbook, as well as Hands-On Graph Neural Networks Using Python.
X | LinkedIn

Viviana Márquez — Developer Relations at Liquid AI
Viviana helps developers build with Liquid's models through technical demos, workshops, and community programs. She is an AI engineer and instructor who has taught more than 4,000 learners worldwide.
X | LinkedIn

Want to showcase your project?

Have you built something using a Liquid decision model? We're looking for production applications, prototypes, experiments, games and creative demos to feature during the showcase.

Selected builders will receive five minutes to demonstrate what they've built, followed by a short audience Q&A.

Apply to showcase: Link
Application deadline: November 1st

Projects must use at least one Liquid decision model. Selected presenters must be able to attend in person in San Francisco.

Questions? Join our Discord or DM our DevRel.

Event details

• 🗓️ November 3rd, 2026

• 🕔 5:30–9:00 PM

• 📍 San Francisco

• 🍱 Food and drinks will be provided

---
By registering for and attending this event, you acknowledge that photos and videos may be taken during the event and may be used by Liquid AI on social media, websites, and other promotional materials.