Evaluating RAG beyond a polished demo
Retrieval quality, groundedness, and the measurements that matter.
SANRID Insights
Thinking clearly about intelligent systems.
Practical writing on AI engineering, agentic systems, automation, and software architecture.
Agents, workflows, and where each belongs
A practical framework for choosing autonomy without adding unnecessary complexity.
Designing APIs for AI-powered products
Contracts, observability, fallbacks, and long-running tasks.
Human oversight is a system feature
Why approval gates and exception handling create better automation.
A production checklist for LLM integrations
Moving from successful calls to dependable products.
Modular architecture for evolving AI systems
Keeping models, retrieval, tools, and product logic loosely coupled.
Start a project