AI Infrastructure Builder — I build the systems that make AI agents work in production: orchestration runtimes, coordination protocols, budget controls, and resilience patterns.
17 years in manufacturing operations. Self-taught engineer. Now building production infrastructure for AI systems.
AI exocortex and multi-agent orchestration monorepo. Three packages: Core (personal AI with learning system and proactive engine), Forge (workflow runtime with 10 agent archetypes, budget enforcement, quality gates, and checkpoints), and Quorum (coordination protocol for parallel agents, live on PyPI).
9,300+ tests · 96% coverage · Python · PyO3 · MIT
Offensive security operations platform. Recon, scanning, vulnerability analysis, and reporting — orchestrated through structured pipelines with full audit trails.
5,000+ tests · Python · MIT
Local-first document intelligence. NER, classification, OCR, full-text search, forensics timeline, and entity graph analysis — all on SQLite with zero cloud dependencies.
391 tests · 99% coverage · Python · FastAPI · MIT
Six CLI tools for AI-assisted development:
| Tool | What | Tests |
|---|---|---|
| promptctl | Claude API toolkit — prompt engineering, code review, doc intelligence | 286 |
| memboot | Zero-infra persistent memory for LLMs — SQLite + TF-IDF vector search | 285 |
| claudemd-forge | CLAUDE.md generator/auditor for AI coding agents | 259 |
| ai-spend | AI API cost aggregator — Anthropic + OpenAI billing in one view | 222 |
| agent-audit | Workflow YAML cost estimator + linter for agent pipelines | 193 |
| mcp-manager | MCP server manager — discover, monitor, manage across IDEs | 139 |
| Project | What | Tests |
|---|---|---|
| Argus Overview | EVE Online multi-boxing tool for Linux (2,500+ downloads) | 1,875 |
| marketing-engine | LLM-powered content pipeline + platform API publishing | 581 |
| TideWise | Fishing forecast engine — solunar, tides, pressure, 7 scoring factors | 373 |
| ai-skills | 66 production skills for Claude Code agents | — |
Animus v2.0.0 shipped — consolidated Gorgon + Convergent into a single monorepo with 9,300+ tests. Current focus: dev tools portfolio and content automation via marketing-engine.
| Tests across active repos | 24,000+ |
| Projects with CI green | 15+ |
| Shipped releases | Argus Overview (2,500+ downloads), convergentAI (PyPI) |
Languages: Python, Rust, TypeScript AI: Claude API, OpenAI, Ollama, multi-agent orchestration, prompt engineering Infrastructure: FastAPI, SQLite, PostgreSQL, Docker, GitHub Actions, Linux Specialty: Production AI systems — budget controls, agent coordination, resilience patterns
- 17 years manufacturing operations (Toyota Production System)
- Self-taught software engineer
- Focus: The boring infrastructure that makes AI actually work in production
Portland, OR · LinkedIn · jamesyng79@gmail.com
