I build autonomous systems that compress months of engineering into days — not by working harder, but by engineering the infrastructure that works without me.
iggy = {
"role": "AI Agent Architect & Systems Engineer",
"background": ["Apple", "AT&T", "Verizon", "Sam's Club"],
"journey": "Embedded → Mobile → Cloud → ML → Agent Systems",
"focus": "Multi-agent orchestration & software factory systems",
"superpower": "Turning frontier AI capabilities into reliable, production-grade tools",
}I've spent my career moving up the stack — from embedded firmware to mobile apps to cloud platforms — and now I'm focused on the next layer: AI systems that build software autonomously. I love the intersection of systems engineering and LLM capabilities, where careful architecture turns unreliable AI into dependable infrastructure.
| Project | What It Does | Tech |
|---|---|---|
| 🛡️ factory-gate | Autonomous structural policy guardrails — verifies safe agent state transitions before target branch ingestion | JavaScript |
| 🔬 circuit-x | Exposes internal feature circuit paths inside deep neural weights and code LLMs for interpretability-driven alignment | HTML · Interpretability |
| 🪟 sliding-window-trim-engine | Efficient sliding window context trimming for LLM token management | Go |
| 🤖 deterministic-agent-supervisor | Deterministic orchestration layer for supervising autonomous agent workflows | Python |
| 🔥 context-forge | Context engineering toolkit for building structured knowledge systems | — |
| 📐 spec-driven-development | A more accessible set of artifacts for Spec Driven Development | ⭐ 1 · MIT |
💡 Every repo here is either a live tool, a proof-of-concept, or a blueprint — designed to compound.
🤖 Multi-Agent Orchestration — Autonomous pipelines where specialized agents plan → execute → verify → iterate across complex workflows without human bottlenecks.
🧩 AI Harness Engineering — Production-grade integration layers connecting LLMs, tools, memory, APIs, and evaluation systems into reliable intelligence.
🏗️ Context Engineering — Structured knowledge systems that make every AI call smarter, more accurate, and more useful than raw prompting.
🏭 Software Factory Systems — AI-accelerated development pipelines enabling a single engineer to ship what once required a team.
🏢 Enterprise AI Deployment — Translating frontier model capabilities into auditable, integrated, durable business value.
| Category | Technologies |
|---|---|
| AI/ML | Claude · OpenAI GPT · Gemini · Llama · Qwen · GLM · Kimi/MiniMax · Multi-Agent Systems · RAG · LoRA · CLIP · Whisper |
| Languages | Python · TypeScript · Go · Swift · Java · Rust · Node · C# · Objective-C |
| Cloud & Infra | AWS · GCP · Azure · Kubernetes · Cloudflare · GitHub Actions · Docker |
| Frontend | React · Svelte · SolidJS · Alpine.js · Vite · Remix |
I'm always happy to chat about agent architectures, context engineering, or how to turn AI research into production systems.
Open to: Collaboration · Consulting · Technical Conversations
The future belongs to builders who engineer their own leverage. ✨