Inference on Tap: Berlin @ NVIDIA GTC

Baseten and CodeRabbit take over Fabrik 23 for an evening on AI inference and the agentic software lifecycle, timed to NVIDIA GTC week in Berlin. The stated ground to cover: fast, scalable and secure inference, fine-tuning open models for narrow high-volume tasks, shipping AI features without shipping surprises, and the risks engineering teams now carry when agents write most of the code. Baseten runs the training and inference platform behind production workloads at Abridge and Cursor; CodeRabbit builds agentic change management and independent validation for code written by people and agents. No speaker list is published — this is a conversation format, not talks. Free, space is limited, and the exact address is shared on registration. Why go: the people who run model deployment and inference performance for a living are in Berlin this week anyway, and this is the room where they compare notes rather than watch a keynote. Best for inference and platform engineers, and anyone owning AI infrastructure cost.