Blog

Research notes, infrastructure war stories, and updates from the workshop.

May 2, 2026 max-enginemodulardgx-sparkinfrastructure

Running Gemma-4-31B on DGX Spark with MAX Engine — What Broke and How We Fixed It

Getting Modular's MAX Engine to serve a 31B parameter model on NVIDIA's smallest Grace Blackwell system.

Read
May 2, 2026 interpretabilitysparse-autoencodersgemma-4datasets

60 Layers of Interpretability: Publishing Gemma-4-31B SAE Features on HuggingFace

3,000 interpreted and verified sparse autoencoder features for every layer of Google's Gemma-4-31B.

Read