Here's a curated list of my writings on software engineering, distributed systems, architecture, and lessons learned along the journey.
Why rigid study routines fail, how to use AI for your awkward questions, and the power of immediate impulse reading over guilt-driven planning.
Why prompt tricks fail, how agent instruction stacks actually work, and the verification contracts you need to keep autonomous coding models aligned.
Transitioning from manual terminal commands to Docker Compose brings declarative consistency to local AI infrastructure. By combining containerized orchestration with Ollama Modelfile definitions, developers can forge specialized AI personas—such as a deterministic developer assistant for clean code generation and a structured personal assistant for workflow management. These tailored local models provide the exact foundation needed for upcoming IDE integrations and custom RAG applications.
Running open-source artificial intelligence locally gives developers complete data privacy, zero API costs, and full offline capability. This beginner-friendly guide demonstrates how to cleanly containerize Ollama using Docker, run popular open-weight models like Llama 3 and DeepSeek, and query your local LLM via REST APIs. It also explores where to discover new models and sets the stage for integrating local AI directly into your code editor.