Writing
Notes from building agentic systems and Web3 interfaces.
Short pieces on the architectural decisions that show up again and again. Field notes, not essays.
- 2026-06NLP basics: NLU, NLG, and core language tasks10 min
A beginner-friendly map of natural language processing, what NLU and NLG mean, and the core tasks that turn text into useful structure.
NLP basics - 2026-06From counting words to embeddings11 min
How NLP moved from one-hot vectors, Bag of Words, and TF-IDF to dense embeddings that represent meaning instead of only counting words.
Embeddings - 2026-06Contextual embeddings and Transformers9 min
Why static embeddings were not enough, how contextual embeddings change with the sentence, and how self-attention helps Transformers model meaning.
Transformers - 2026-06Intent classification is the routing layer for LLM apps10 min
How to turn natural language into structured intents, metadata, validation, and workflows your product can actually execute.
Intent routing - 2026-05From keyword search to agentic AI6 min
How retrieval evolved from exact word matching to agents that decide what to search for, and why that shift is the harder problem now.
Retrieval - 2026-05Agent memory is not agentic storage7 min
Two related ideas that get fused in production code. A simple separation that makes agents easier to debug, eval, and scale.
Agent memory