Forward Deployed Software Engineer | AI Enablement | Operational Intelligence
17 years in manufacturing operations. I now build automation, AI workflows, and operational systems to replace tribal knowledge and firefighting.
Multi-agent AI orchestration framework
Coordinate specialized AI agents (Planner → Builder → Tester → Reviewer) across enterprise workflows. FastAPI backend, React dashboard, token/cost tracking. 85K+ LOC, 3600+ tests.
Demonstrates: AI workflow design, production instrumentation, agent coordination
Personal AI exocortex framework
Local-first memory system with vector storage, multi-model reasoning, and self-learning with guardrails. 18K LOC, 333 tests.
Demonstrates: AI memory architecture, safety guardrails, local-first design
Professional multi-boxing tool for EVE Online (Linux)
Real-time 30 FPS window previews, team management, auto-tiling layouts, EVE settings sync. 2,500+ downloads.
Demonstrates: Desktop application development, performance optimization, user-facing tools
Fantasy sports decision engine with AI
iOS app with Claude AI integration for lineup optimization, injury analysis, and game predictions. Token tracking, budget monitoring, real-time data sync.
Demonstrates: Mobile AI integration, real-time sports data, subscription systems
Top-down arcade shooter in the EVE Online universe
Rust/Bevy rewrite with Elder Fleet Liberation campaign, 3 game modes, procedural audio, WASM web builds.
Demonstrates: Game development, Rust/ECS architecture, cross-platform builds
| Project | Description |
|---|---|
| EVE Gatekeeper | Fleet intel, route planning, jump bridge networks — backend + mobile + desktop |
| Argus Overview | Multi-boxing tool for Linux with window previews |
| EVE Sentinel | Alliance intel and recruitment risk scoring |
| EVE Quartermaster | Mobile companion — assets, routes, market |
| Project | Description |
|---|---|
| LikX | Screenshot capture — X11/Wayland, OCR, annotations, cloud upload |
| G13 Linux | Logitech G13 driver with macros, RGB, LCD display |
| RazerControls | Razer peripherals control center for Linux |
| SteamProtonHelper | Proton/Wine repair utility with CLI + GUI |
| Domain | Technologies |
|---|---|
| Languages | Python, Rust, TypeScript, Swift |
| AI/ML | Claude API, OpenAI, LangChain, prompt engineering |
| Backend | FastAPI, Flask, Firebase, PostgreSQL, SQLite |
| Frontend | React, Next.js, Streamlit, PySide6/Qt |
| DevOps | Docker, GitHub Actions, Linux administration |
| Specialty | Operational systems, multi-persona design, offline-first architecture |
Every AI system I build follows these principles:
| Layer | Implementation |
|---|---|
| Structured Outputs | JSON schemas, Pydantic validation, typed responses |
| Safety Rails | Input sanitization, content filtering, rate limiting |
| Retry/Fallback | Exponential backoff, graceful degradation, circuit breakers |
| Telemetry | Token usage, latency tracking, cost monitoring |
- 17 years in manufacturing operations (Toyota)
- Current: AI Enablement & Workflow Analyst — converting ad-hoc AI experiments into documented, repeatable workflows
- Focus: Forward Deployed Software Engineering, Solutions Engineering, AI Enablement
- Location: Portland, OR
- Email: jamesyng79@gmail.com
- LinkedIn: linkedin.com/in/james-young-3b77b3120
I build systems that give operators the visibility they need to make better decisions faster.


