AI ENGINEERING
Modern product engineering is being transformed by AI-driven developers who understand both code and product vision. These engineers can effectively translate product requirements to AI systems, then review and optimize the generated code for quality, consistency, and performance. This approach eliminates common mistakes, accelerates development cycles, and ensures the final product meets both technical excellence and business objectives.
9 min read
Claude Code for Engineering Teams: Building a Shared CLAUDE.md That Actually Stays Current
Running Claude Code solo is easy. Running it across a 5-10 person team without everyone developing conflicting workflows, duplicate context, and a CLAUDE.md that nobody trusts is the hard part. Here's the org design, not the technical architecture.
10 min read
CI/CD for AI: Running Prompt Regression Tests and LLM-as-a-Judge Gates in GitHub Actions
GitHub shipped agentic workflows in February 2026. But your existing CI/CD isn't equipped for AI outputs - deterministic tests break on probabilistic outputs. Here's how to run prompt regression tests and LLM-as-a-Judge evaluation gates that actually catch regressions.
10 min read
Which MCP Server Should You Build? A Decision Guide for the 5 Real Patterns
13,000+ MCP servers exist but the tutorials don't tell you which pattern to use or why. stdio vs HTTP vs .mcpb vs npm package vs raw script - here's the decision guide nobody has written yet, with honest tradeoffs for each.
8 min read
Claude Code Hooks: The Reliability Layer Your CLAUDE.md Can't Replace
CLAUDE.md instructions get followed 70-90% of the time. Claude Code hooks run 100% of the time, every time, automatically. Here's what hooks actually do, the patterns that matter in production, and why this is the most underused feature in Claude Code.
6 min read
AGENTS.md vs. CLAUDE.md: What Claude Code Actually Reads (And the Workaround for Both)
Claude Code reads CLAUDE.md. Other agents read AGENTS.md. Here's what the two standards actually do differently, why Claude Code doesn't support AGENTS.md yet, and the exact workaround to manage both without duplicate work.
9 min read
Context Engineering for Claude Code: Why Most Agent Failures Are Context Failures
Context engineering is the discipline of designing what information your AI agent has, in what format, at what time. For Claude Code specifically: CLAUDE.md, tool definitions, memory compaction, and project structure are all context engineering decisions - here's how to approach them as a system.
12 min read
Using Claude Locally in 2026: Desktop, Code, and Fully Offline - What Actually Works
Can you run Claude locally or offline in 2026? Here's what actually works: Claude Code with Ollama (3 env vars), Claude Desktop's new third-party gateway, MCP bridges, and fully offline alternatives when you need zero cloud dependency.
10 min read
Running Claude Code with a Local LLM in 2026: No Proxy Required
How to run Claude Code with a local LLM using Ollama, LM Studio, or oMLX - no proxy required. Model recommendations by hardware, 2026 setup commands, and the fastest backends available today.
10 min read
The Revolt Against Bloat: What Happens When a Builder Gets Tired of Bad Software
From building my own ATS, documentation tool, and Chrome extension to watching a solo dev replace Logitech's entire software suite in a week. AI is collapsing the cost of building what commercial software should have been.
14 min read
From Spec to Skill: The Mental Model That Made Claude Code Click
After hitting the 'context wall' repeatedly, one mental shift changed everything: treat Claude Code like an OS, not a chatbot. Here's the vocabulary - workflow, skill, agent, sub-agent - and why it matters for how you structure every session.
14 min read
Your AI Tooling Budget Is Already Unsustainable - Here's What to Do About It
Uber burned through its entire annual AI budget in four months. Most teams don't notice the problem until it's already a crisis. Here's how to diagnose token waste and fix it before your CFO asks.
14 min read
From $500/Day to $0: Investigating Local LLM Inference for Heavy Claude Code Users
I'm spending serious money on Anthropic every week running Claude Code across multiple parallel projects. Before I pull the trigger on a Mac Studio, I'm sharing my full analysis - the cache hit math, the quality tradeoffs, and why the answer isn't as simple as it looks.
12 min read
Prompt Injection: The Hidden Risk in Copy-Paste AI Culture
Copy-pasting AI prompts from the internet is convenient - but dangerous. Learn how malicious prompts can steal your secrets, leak private data, and what you can do to protect yourself.
8 min read
Chrome DevTools MCP: AI Agents Can Finally See What They Build
AI coding agents write code fast, but they can't see what breaks in the browser. Chrome's new DevTools MCP server changes that—giving AI the debugging tools developers rely on every day.
9 min read
The Real Risks of Relying on AI to Speed Up Coding — And How Teams Win When They’re Smart
AI can speed delivery but often hides costs. This article outlines the most common pitfalls in AI-assisted coding—missed edge cases, security gaps, and technical debt—and shows practical workflows teams use to ship responsibly.
10 - 14 min read
AI‑Native Development: How I Combined Claude Code, Zencoder, and Codex‑Style Parallel Agents
A practical blueprint for an AI‑native engineering pipeline that pairs repo‑aware generation and validation with parallel agents. What changed, how it works, and how to implement it without breaking your team’s workflow.
8 min read
How Wispbit MCP Solved My AI Code Quality Problem
AI agents write a lot of code, but most of it needs work. Here's how Wispbit's MCP integration turned my AI coding workflow from chaotic to consistent.
4 min read
Running Claude Code Locally Just Got Easier with ollama-code
A lightweight wrapper around Ollama that mimics Claude Code's development experience using open models like codellama, deepseek-coder, or codestral locally.
12 min read
Spec-Driven Development: Keeping the Vibe, Adding a Plan
Discover how spec-driven development combines the creative flow of AI pair programming with the structure needed for production-ready applications. Learn why vibe coding breaks down at scale and how specifications can guide AI to build better, more maintainable software.
12 min read
Extending AI's Memory: RAG, CAG, Long Contexts and Vector Search
Learn how RAG, CAG, vector databases, and long context windows are revolutionizing AI systems. Discover strategies to reduce hallucinations, implement efficient retrieval, and build more reliable AI applications with grounded data.
15 min read
Building an AI-Powered Code Review System: A Technical Deep Dive
Learn how to build CodeBot, a Laravel-based system that automatically reviews GitHub pull requests using Google Gemini and Anthropic Claude. This technical deep dive covers architecture decisions, implementation challenges, and production-ready patterns for AI integration.
12 min read
Zencoder vs Cursor vs Claude Code: Why I Made the Switch
Zencoder vs Cursor vs Claude Code - an honest comparison from someone who used all three daily. Here's what Repo Grokking and custom agents do that the others can't, and when Zencoder actually wins.
12 min read
AI-Driven Development: Streamlining Product Engineering with MCP, Linear, and Cursor
Learn how integrating Linear's MCP server with Claude and Cursor transforms product development workflows. See how to generate architectural guidelines automatically from tickets, enabling AI to understand context and speed up development cycles while maintaining code quality.
8 min read
Comprehensive guide for product engineers on building Success
A comprehensive guide for products. Learn 5 key strategies including problem-first thinking, data foundations, explainable systems, and thoughtful human-AI relationships. Avoid common pitfalls engineers on building successful AI-driven products. Learn 5 key strategies including problem-first thinking, data foundations, explainable systems, and thoughtful human-AI relationships. Avoid common pitfalls and create products that deliver real user user value.
10 min read
Running Claude Code with a Local LLM: A Step-by-Step Guide
Learn how to set up Claude Code with a local large language model (LLM) using the code-llmss project. This guide walks you through installation, configuration, and real-world use cases for developers who want AI-powered coding assistance without relying on cloud-based services.
7 min read
Balancing AI Capabilities and User Experience
Learn how to create AI products that are both technically impressive and user-friendly. This guide offers practical strategies for finding the perfect balance between advanced AI capabilities and intuitive user experiences.