# ramenbuild

> Workflow-first AI MVP agency for seed-stage founders. We design and build agentic workflows that reason, retrieve, decide, and act across users' real jobs.

ramenbuild helps seed-stage founders scope, design, and build the right first version of an AI-native product. We specialise in agentic workflow design — multi-step systems where LLMs use context, tools, and decision logic inside real operational workflows, not just generate a response.

## Pages

- [ramenbuild homepage](/index.md): Full site content — services, process, philosophy, and contact.

## Blog

- [Blog index](/blog): All articles — agentic workflow design, AI architecture, and building AI-native products.

## Articles

- [Agentic workflows](/agentic-workflows/index.md): What agentic workflows are, how they work, when to use them, and how to build and evaluate them. Covers architecture patterns, real-world use cases, build approach (low-code vs. custom), evaluation, common pitfalls, FAQ, and glossary.
- [Recruitment agent: talent matching workflow](/recruitment-agent/index.md): A browser-based agentic workflow demo covering structured data extraction from job posts and CVs, retrieval, and personalised job matching — all running locally with a quantised LLM via WebLLM.