I approach software like a laboratory process: experimenting, failing fast, iterating, and occasionally landing on a deterministic build worth keeping. My repositories are working essays, not polished commercial showcases.

• Determinism First: If a task can be solved with a predictable script, never waste compute on a LLM.
• Privacy & Local Execution: Data belongs to the user. My tools prioritize local runtimes (Docker, local models, native scripts).
• Frugality & Zero Pollution: Minimal dependencies, ephemeral containers, and low cognitive overhead.
• Empirical Rigor: Continuous iteration and pragmatic architectures over over-engineered solutions.

• 📄 drop-to-ocr — Zero-friction, drag-and-drop Windows batch wrapper around OCRmyPDF running in ephemeral Docker containers.
• 🤖 agent-builder — Framework for local agent execution and pragmatic automation architectures.
• ⚙️ qse-studio — Systems and workflow management tools focused on quality and operational rigor.

• 🌐 Website: nicolas-hue.com
• 💼 LinkedIn: linkedin.com/in/nicolas-hue91