Own your inference: What it Takes to run Open Models for your Team

PowerTalk

About the session

Coding agents are essential to how a lot of us work now, and almost none of us own the stack they run on. We use them through closed APIs and subscriptions, which means someone else decides the price, the quality, and whether the model stays available at all. Open coding models are a real alternative today, and the best of them are now very close to frontier models.

This session is a practical guide to running open coding models for your team. We walk through what actually matters when you self-host: choosing the right model and hardware for your budget, quantization, serving engines, agent harnesses, and the per-developer economics of running your own stack. Along the way, we share what we learned putting these models to work on real coding tasks, including the configuration choices that quietly make or break quality.

You will walk away with a clear playbook: which model fits your budget, how to set it up right, what it actually costs, and how to give your team coding agents on a stack you fully control.

Speaker

Download Brochure