Exploring transformer architecture patterns for production-grade semantic search systems, including custom embedding pipelines and sub-100ms query optimization.

Reflections on building and leading remote engineering teams across Asia-Pacific, the tools that actually matter, and why async communication is a superpower.

Everything I learned fine-tuning large language models for production use cases — from data preparation to evaluation, with real numbers and hard-won lessons.

Lessons from building an orchestration platform for autonomous AI agents — task decomposition, memory sharing, and why most multi-agent demos fail in production.