Tech blogWe don't just build. We think out loud.

Misaligning LLMs with supervised finetuning

Evaluating RAG systems with synthetic data and LLM judge

Few-shot information extraction from PDF documents