$ cat ~/notes
Field notes from building products end-to-end - fullstack TypeScript, self-hosted infra, and the occasional foray into local LLM inference. Mostly things I wish I'd known six months ago.
$ cat post.md ───────────────── # what-is-rag-how-retrieval-augmented-generation-actually-works tags: [RAG, LLM, AI Engineering, Vector Databases, Machine Learning, System Design] read: ~1min
What is RAG? How Retrieval-Augmented Generation Actually Works
A practical breakdown of Retrieval-Augmented Generation — why LLMs hallucinate, how RAG grounds answers in real documents, and the architecture decisions that make or break a RAG system in production.
Why Distributed Systems Break at the Edges: Retries, Events, and the Outbox Pattern
The difficult part of distributed systems isn't sending a message. It's handling what happens when databases, queues, and services fail at different times. A practical breakdown of retries, idempotency, dead-letter queues, and the Outbox Pattern.
Low-Level Design Isn't About UML: It's About Designing Code That Survives Change
Low-Level Design is less about drawing class diagrams and more about managing complexity. A practical look at interfaces, abstractions, dependencies, and designing systems that remain maintainable as requirements change.
Why Solana Feels Different to Build On: Understanding Its Architectur
Solana's performance comes from architectural decisions that differ from traditional blockchain systems. This article explains the ideas behind parallel execution, accounts, transactions, and high-throughput design.
What Happens When AI Agents Get Their Own Wallets?
AI agents can already reason and use tools. Giving them wallets introduces a new capability: the ability to pay, transact, and interact directly with on-chain systems.
AI Agents Aren't Just Chatbots: How Agentic Systems Actually Work
AI agents are moving beyond question-and-answer systems. This article breaks down how agents use reasoning, tools, memory, and feedback loops to take actions and complete real workflows.
I Built an AI Coding Agent in Rust — Here's What Actually Happens Under the Hood
AI coding agents are more than chatbots with access to a terminal. Here's how I approached building Codey in Rust, from context and sessions to tools, MCPs, Skills, and plugins.
New posts hit your reader instantly — no email, no tracking.