The Inference Engineering Masterclass — Philip Kiely & Ali Taha, Baseten
Baseten raised $13B Series F as inference engineering emerges as a critical AI discipline
“How do you turn those weights from training into a product that is fast, reliable, and affordable at scale?”
Baseten has raised a $13B Series F, cementing its status as an AI infrastructure decacorn at the peak of what the authors call 'the Inference Inflection.' The piece frames inference engineering as a distinct discipline that barely existed three years ago but now addresses the core production challenge of converting trained weights into fast, reliable, affordable products. A notable technical finding: quantizing a GLM-5.2 model preserved benchmark quality while increasing throughput 20%, because quantization errors across layers can cancel out.