Software engineer building AI agent systems that run in production, not demos.
9+ years of shipping software. Currently focused on AI-enabled engineering — designing multi-agent systems, developer tooling, and infrastructure that lets small teams move like large ones.
I run a 5-agent AI system (TitanAgent) in production 24/7. It does content research, code review, and deployment coordination across Discord, GitHub, and AWS. I also built a config-driven content engine on ECS + K3s + Supabase that handles the entire publish pipeline without human intervention.
I write about what I build: architecture decisions, trade-offs, things that broke in production, and what I'd do differently. No tutorials, no hype — just notes from someone shipping AI systems daily.
Multi-agent architecture, AI development workflows, prompt engineering at scale, and production reliability for LLM systems.
CLI tools, config-driven architecture, IaC, observability, and building internal platforms that reduce friction.
System evolution logs, design decisions, trade-offs, and production incidents. The messy reality of shipping.
Check out the projects page for architecture details, or the blog for ongoing build logs.