The lab notebook
Field notes on AI engineering
824 articles · page 20 of 28
OpenAI Codex vs GitHub Copilot Workspace: A Comparison
A practical, no-hype comparison of OpenAI Codex and GitHub Copilot Workspace for engineers deciding which agentic coding tool fits their workflow.
Pair Programming with Claude Code: Best Practices
Practical workflows for pairing with Claude Code — from context management to review discipline — so you ship faster without losing control of your codebase.
Parent-Document Retrieval: Solving the Small-Chunk Context Problem
Small chunks retrieve well but read poorly — parent-document retrieval fixes the mismatch by searching one thing and returning another.
What a Production-Grade Capstone Demo Actually Looks Like
Most capstone demos are notebooks pretending to be products — here's the checklist that separates a real production demo from a screen recording of localhost.
Prompt Regression Suites: Structuring Tests Like a Codebase
Learn how to structure prompt regression suites with the same rigor as software tests, so prompt changes stop breaking production silently.
Publishing Your MCP Server: Packaging and Distribution
A practical guide to packaging, versioning, and distributing your MCP server so other developers can install and trust it.
Building a RAG Chatbot with Streaming Responses
A practitioner's guide to building a RAG chatbot with streaming responses, from retrieval pipeline to token-by-token UI updates.
Chunk Size and Overlap in RAG: Finding the Sweet Spot
A practical guide to picking chunk size and overlap in RAG pipelines, with code, trade-offs, and debugging tips.
RAG Chunking Strategies Compared: Fixed, Recursive, Semantic and Agentic
A practitioner's comparison of fixed, recursive, semantic, and agentic RAG chunking strategies, with code and guidance on when to use each.
RAG Cost Optimization: Cutting Your Vector DB and LLM Bill
A practitioner's guide to cutting RAG costs across embeddings, vector storage, retrieval, and generation without wrecking answer quality.
RAG Data Pipeline: From Raw Documents to Searchable Chunks
A practitioner's guide to building a rag data pipeline that turns messy PDFs and docs into clean, searchable chunks.
RAG Deployment Checklist: 12 Things to Verify Before Launch
A practical pre-launch checklist covering chunking, retrieval, evaluation, and cost so your RAG system doesn't fall apart in production.
RAG Evaluation Metrics Beyond Ragas: A Broader Toolkit
Ragas is a great starting point, but shipping RAG in production needs a wider set of metrics, harnesses, and habits.
RAG for Codebases: Building a Chatbot That Understands Your Repo
A practical walkthrough of building a code-aware RAG chatbot, from chunking strategy to retrieval tricks that actually work on real repos.
RAG for Customer Support: Building a Support Bot That Cites Sources
A practical guide to building a RAG-powered support bot that cites its sources, with retrieval, prompting, and evaluation code.
RAG for E-Commerce: Product Search That Understands Intent
A practical guide to building RAG-powered product search that understands shopper intent instead of just matching keywords.
RAG for Internal Wikis: A Confluence/Notion Knowledge Base Bot
A practitioner's guide to building a RAG bot over Confluence or Notion, covering sync, chunking, retrieval, and the failure modes nobody warns you about.
RAG for Legal and Compliance Documents: Accuracy Requirements
A practical breakdown of the accuracy, citation, and audit requirements that separate a demo-quality legal RAG system from one you can actually ship.
RAG for Onboarding Docs: Turning Your Handbook Into a Chatbot
Learn how to convert a static employee handbook into a searchable, accurate RAG-powered chatbot new hires actually use.
RAG for PDFs: Handling Tables, Images and Multi-Column Layouts
A practitioner's guide to building rag for pdfs that actually survives tables, scanned images, and multi-column academic layouts.
RAG Hallucination: Why It Happens and How to Reduce It
RAG hallucination doesn't disappear just because you added retrieval — here's why it still happens and the concrete fixes that reduce it.
RAG Latency Optimization: Making Retrieval Feel Instant
A practical guide to cutting RAG pipeline latency at every stage, from embedding to reranking to generation, with code you can ship today.
RAG Observability: Tracing Every Retrieval and Generation Step
Learn how to trace retrieval, ranking, and generation in RAG pipelines so failures stop being invisible.
Building a RAG Pipeline with LlamaIndex: A Complete Walkthrough
A hands-on walkthrough for building a production-ready RAG pipeline with LlamaIndex, from document loading to query-time reranking.
RAG Query Rewriting and HyDE: Improving Retrieval Before You Search
Learn how query rewriting and HyDE fix bad retrieval in RAG systems by transforming user queries before they ever hit your vector store.
RAG Security: Preventing Prompt Injection Through Retrieved Content
A practical guide to rag security prompt injection risks in production systems, with concrete mitigations, code patterns, and detection strategies.
Handling Stale Data in RAG: Cache Invalidation for Knowledge Bases
A practical guide to detecting and fixing RAG stale data problems with embedding versioning, TTLs, and event-driven invalidation.
RAG vs Long-Context LLMs: Do You Still Need Retrieval in 2026?
A practitioner's honest comparison of RAG vs long context in 2026, with real tradeoffs, code, and a framework for choosing.
RAG Without a Vector Database: Simpler Alternatives That Work
You don't need Pinecone or a cluster to ship RAG — here are simpler retrieval alternatives that work for most real-world document sets.
Ragas in CI/CD: Automating RAG Regression Tests
Learn how to wire Ragas metrics into GitHub Actions so every RAG pull request gets scored for faithfulness and relevancy before it merges.