Building systems that last. Technical deep dives, architecture decisions, and engineering practices from production experience.
System design, technical patterns, avoiding the rewrite trap, and making technology choices that scale.
Retrieval gives agents information. Context gives them decision logic, constraints, and workflows. This deep dive explains why RAG cannot solve decision-making and why the context layer becomes essential infrastructure.
Companies are racing to add AI everywhere. But most are strapping rockets to donkeys and calling it innovation. Real transformation requires rebuilding systems, not just adding layers on top.
Stablecoins promised safety, but without compliant yield, savers in inflation-hit markets lose value. CTO lessons on why DeFi lending survived, CeFi failed, and where safe yield might emerge.
Most blockchain pilots failed, but the survivors now move real value. Stablecoins, DeFi collateral, and corruption-proof records show where blockchain’s design choices actually deliver.
Many teams are tempted to rebuild a clean V2 from scratch, but this often fails because it discards history, tribal knowledge, and unwritten context. A more effective path is gradual migration with strong boundaries rather than extremes like full microservices. With AI reshaping the way we build, the challenge is knowing where it can augment workflows without replacing engineering discipline. The teams that thrive will move deliberately, applying clarity and judgment to every step.
The phalanx offers a powerful metaphor for engineering culture. It is not about individual heroes, but trust, discipline, and leaders who stand unshielded for their teams. Phalanx Engineering is a way of thinking about resilience that I’ve applied and refined in practice.
