Local LLMs & Agentic Coding: The end of "Boilerplate" for Senior Devs? I’ve been diving deep into Local LLMs and Agentic coding workflows lately, and the shift in productivity is staggering.

As a Senior Developer with 10+ years in the game, I’ve seen every "productivity silver bullet" come and go. But this feels different. I’m currently experimenting with local models to handle the repetitive heavy lifting—generating PHPUnit tests for my Laravel controllers.

The realization? AI agents aren't replacing the Senior Architect; they are amplifying us. They handle the "syntax," while I focus on the system design, API security, and clean MVVM architecture.

However, there’s a catch: An agent is only as good as the prompt and the context you give it. If your codebase is a mess, the AI will just help you write "messy code" faster.

My current Local LLMs stack: Ollama, OpenCode, and Qwen3.5 9B

The Question: Are you using AI agents to write your code, or to architect it? And for my fellow devs—are you running your models locally yet, or sticking to the cloud?

Let’s talk shop in the comments!