kefa

Indie builder exploring practical LLM applications beyond prompts - focusing on reliability, validation, and real-world behavior of AI systems. I write about what breaks when you actually ship LLM too