No fluff. Things we've learned the hard way — from RAG pipelines to agent orchestration, evaluation harnesses to LLM cost control, OJK compliance to vector databases. Some posts are AI-drafted and human-reviewed; we always disclose.
One email a month. Practical AI delivery notes. No vendor pitches.
Menilai output AI dengan perasaan tidak akan bertahan di produksi. Pelajari cara membangun harness evaluasi yang objektif, dapat diulang, dan menangkap regresi sebelum pengguna.
9 min · Jun 2026 · Super Admin
