https://suraj.io/tags/attention/ Recent content in Attention on Suraj Deshmukh https://suraj.io/images/papermod-cover.png https://suraj.io/images/papermod-cover.png Hugo en Sun, 02 Aug 2026 10:00:00 +0530 https://suraj.io/post/2026/inference-engineering-visualizations/ Sun, 02 Aug 2026 10:00:00 +0530 https://suraj.io/post/2026/inference-engineering-visualizations/ A set of interactive, self-contained visualizations that make the core concepts of inference engineering — dimensionality, attention, MoE, and the roofline model — tangible.