Train Custom Models with Fireworks & Microsoft + NYC Happy Hour ⭐

Join Fireworks and Microsoft September 2nd for an invite only hands-on workshop: get your hands on Fireworks and train a real model yourself.

Across the industry, model reliability is the number one reason agents fail to complete a task. Frontier models demonstrate capability but then can’t constantly deliver at production grade scale.

This lab explores how post-training drives reliable model behavior, unlocking performance on smaller, faster models families. We’ll cover when to use post-training, the trade-offs involved, and a hands training session that leaves you with a customized model that you can upload to Azure Foundry.

Agenda

5:00pm - Doors open: Networking over bites and drinks

6:00pm - Opening remarks on how Specialized Intelligence is transforming product development

6:30pm - A hands-on lab where we set a baseline, train an open model, and measure the results. Then we help you upload your model to Azure Foundry for production-ready serving.

8pm - Doors close

Who should attend

This session is designed for engineers and product managers who want a practical understanding of how to build specialized models. Some familiarity with machine learning concepts is helpful, though deep expertise is not required.

What to bring

A laptop, so you can follow along with the hands-on portions.

About us:
Fireworks is the platform for specialized intelligence, enabling companies such as Doximity, Revolut, and Shopify to train and serve models tailored to their own data, workflows, and use cases. Founded by the team behind PyTorch, and backed by AMD, Atreides, Benchmark Capital, Index Ventures, Lightspeed, NVIDIA, Sequoia Capital, and TCV, Fireworks provides hundreds of state-of-the-art open models in text, image, embedding, and multimodal formats globally.