Soumik Paul·Sep 24Decoding Transformers, One Equation at a Time: Position-wise Feed-Forward NetworksPart 3 of a learning-in-public series on the Transformer architecture
Soumik Paul·Sep 8Decoding Transformers, One Equation at a Time: Multi-Head Attention and Where Attention Is UsedPart 2 of a learning-in-public series on the Transformer architecture
Soumik Paul·Aug 31Decoding Transformers, One Equation at a Time: Scaled Dot-Product AttentionPart 1 of a learning-in-public series on the Transformer architectureA response icon1A response icon1
Soumik Paul·Jun 26Design Patterns Are Not Dead. Your AI Agent Just Forgot to Invite Them.Before your AI agent learn another trick, teach it some software engineering manners. A practical, playful look at why design patterns…