FeaturedRAG on a CV-Sized Corpus: How This Site Answers Questions About Me
How I built a grounded RAG chatbot for a portfolio site — whole-item embeddings, hybrid Atlas search, Gemini answers, and defense-in-depth grounding on a CV-sized corpus.
Knowledge Base
Technical Articles
Exploring the intersections of high-performance systems engineering, AI-native architectures, and the future of technical software design.
FeaturedFrom Feature Folders to Hexagonal Layers: A Mid-Project Refactor
The backend did not start hexagonal. How eighteen feature folders became four layers in one 148-file commit — port ownership, handler splitting, transitional facades, and what the discipline costs day to day.
Hexagonal Architecture and DDD on a Chat-First Portfolio
How ports-and-adapters and DDD-lite tactical patterns shape this site's Rust backend — the layer map, the ten ports, and the full pipeline from embedding to grounded chat, traced through the architecture.
What Actually Went Wrong: Lessons from Building a Grounded Chatbot
Security, grounding, streaming, proxies, and intent routing — the real difficulties from a month of shipping RAG on this portfolio, and how each one was fixed.
Agentic AI Without the Agent Loop
Why this portfolio chatbot feels agentic — intent routing, catalog-in-context, and navigation-as-answer — without LLM tool calling or a multi-step agent loop.