What I Build. What I Write About.

I don't optimize for frameworks, and I don't track vanity metrics. I build useful things, fast.

On my Substack, I document the exact engineering principles, architectural patterns, and execution strategies used to ship these projects under real-world constraints.

If you are navigating the gap between high-tech complexity and actual business results, these are the core domains I break down:

1. AI Engineering & Production Workflows

Moving past the wrapper hype to build reliable, production-grade AI infrastructure.

  • Deterministic Output: Strategies for grounding LLMs and handling structured data validation.
  • Agentic Orchestration: Designing autonomous workflows that don't loop endlessly or waste compute.
  • Cost & Energy Efficiency: Building minimal, energy-conscious background layers and architectures.

2. Model Context Protocol (MCP) & Knowledge Systems

Bridging the gap between developmental runtime environments and local cognitive architecture.

  • Obsidian MCP: Extending local, markdown-based knowledge graphs (PARA/Zettelkasten) into executable tools for LLMs.
  • Contextual Framing: How to structure deep technical context so AI agents can make better decisions with less token waste.

3. Open Source & Pragmatic Tooling

Building minimalist, high-performance utilities from source, focusing on security, speed, and zero bloat.

  • Rust & Minimal Daemons: Event-driven tools (like structured NDJSON file watchers) built for speed.
  • Distroless & Containerization: Hardening CI/CD pipelines by stripping away everything but the executable.

4. Startup Execution & Real-World Shipping

The mechanics of taking an idea from a scrappy MVP to a functional system without burying yourself in technical debt.

  • Velocity Over Perfection: Tactical lessons on narrowing project scope to ship value on day one.
  • Legacy Rescues: The unglamorous, highly profitable art of stepping into broken enterprise systems or WordPress wrecks and stabilizing them.

Outcomes > Frameworks.

Value > Complexity.

Judgment > Code Volume.

If this approach to engineering resonates with you, you should join the publication. No growth hacks, no marketing noise. Just the raw logs of a builder working at the edge of AI and software architecture.

Browse by Tag