AI Agent Developer with hands-on experience across construction, energy consulting, and manufacturing. I don't just build prototypes—I ship production systems that generate revenue and solve real operational problems.
Current Focus:
- Claude Agent SDK & Multi-Agent Orchestration
- MCP Server Development
- Production AI Systems with Measurable Business Impact
| Metric | Value |
|---|---|
| Production Code | 385,000+ lines |
| Production Systems | 5 deployed |
| Git Commits | 200+ |
| AI Tools Built | 15+ specialized agents |
| Monthly Recurring Revenue | $500 |
| Revenue Pipeline | $2,145+ |
| Cost Efficiency | 918x |
| Development Value | $458,900 in 65 days |
Last updated: February 2026
Status: 🟢 Production | Revenue Generating:
Production AI system for EV charging grant research serving Law Ventures LTD - a ComEd-certified EV charging and energy efficiency contractor in Chicago. Features AVA conversational agent with 12 specialized tools covering grant eligibility (200K+ programs indexed), NEVI compliance verification, and automated funding calculations.
Tech Stack: Claude API, Node.js, TypeScript, Railway, PostgreSQL, n8n, WebSocket
Impact:
- Reduced grant research time: 40 hours → 15 minutes
- Increased program identification: 3-5 → 20+ grants per query
- Client ROI: 10x cost savings vs manual research
View Architecture | See Code Patterns
Status: 🟢 Production
Real-time multi-agent orchestration platform managing 15+ specialized subagents including: grant researcher, compliance expert, form filler, agent developer, workflow builder, documentation writer, autonomous executor, session manager, streak tracker, roadmap coordinator, and more.
Tech Stack: Claude Agent SDK, FastAPI, Python, MCP Servers, WebSocket, HTTPS/TLS
Features:
- WebSocket streaming for real-time updates
- API key authentication
- Secure HTTPS deployment
- 6-module Claude Agent SDK tutorial progression
View Architecture | See Code Examples
Status: 🟡 95% Complete | Launch: January 2026
Consumer-facing SaaS application for AI agent services with React frontend, FastAPI backend, and 4-tier pricing model (Free/$99/$499/Enterprise).
Tech Stack: React, Vite, FastAPI, JWT Authentication, Stripe, PostgreSQL
Features:
- 7 React components (4,339 lines)
- 8+ FastAPI endpoints
- Full SEO implementation
- Beta feedback collection system
Status: 🟡 MVP Ready | Client: Maine Scientific
FDA/GMP/HACCP compliance assistant for Maine Scientific - a GMP consultation and formulation services company. Local Ollama deployment ensures complete data privacy with zero API costs.
Tech Stack: Ollama, PowerShell, Python, PostgreSQL, 21 CFR, HACCP
Features:
- Local AI model (994MB, privacy-first)
- PowerShell automation module
- 25-table compliance database
- FDA recall searches & 21 CFR lookups
View Architecture | See PowerShell Patterns
I created a comprehensive guide for optimizing Claude Code's CLAUDE.md configuration file based on Anthropic's best practices.
╔═══════════════════╦═══════════════════╦═══════════════════╦═══════════════════╗
║ 71% ║ 168 ║ 4 ║ ~3,100 ║
║ Size Reduction ║ New Line Count ║ Modular Rules ║ Tokens Saved ║
╚═══════════════════╩═══════════════════╩═══════════════════╩═══════════════════╝
| Metric | Before | After |
|---|---|---|
| Total Lines | 586 lines | 168 lines |
| Estimated Tokens | ~4,500 | ~1,400 |
| Modular Rules | 0 files | 4 files |
| Resource | Description |
|---|---|
| Optimization Guide | Step-by-step guide to reduce token usage |
| Modular Rules Templates | Reusable .claude/rules/ examples |
| Visual Report | Interactive before/after comparison |
- Keep
CLAUDE.mdunder ~200 lines (~2,000 tokens) - Use
.claude/rules/for modular, topic-specific instructions - Move documentation to
docs/, not inline - Use
@pathimports to reference external files
- Agent Frameworks: Claude Agent SDK, Claude Code, MCP Server Development
- APIs: Anthropic API, Ollama (local models)
- Techniques: RAG Systems, Vector Embeddings, Prompt Engineering, Multi-Agent Orchestration
- Languages: Python, TypeScript, JavaScript, Node.js
- Frameworks: FastAPI, Express
- Real-time: WebSocket, Server-Sent Events
- Databases: PostgreSQL, SQLite, Supabase
- Deployment: Railway, Docker
- CI/CD: GitHub Actions, Railway Auto-Deploy
- Security: HTTPS/TLS, API Key Auth, JWT
- Monitoring: Health checks, logging, alerts
- Workflow: n8n Cloud, PowerShell scripting
- Integration: Slack, Airtable, Webhooks
- APIs: RESTful design, OpenAPI/Swagger
- React, Vite, HTML5/CSS3, Responsive Design, Dashboard UI/UX
- Energy Sector: ComEd certified, NEVI/CFI grants, EV charging infrastructure, energy auditing
- Regulatory Compliance: FDA (21 CFR), GMP, HACCP, OSHA safety standards
- Grant Programs: Federal grants (Grants.gov), state energy programs, utility incentives
- Event Production: Large-scale event management, vendor coordination, SOP development
├── docs/ # Resume, CV, credentials
├── projects/ # Project architecture & documentation
│ ├── evlv-grant-platform/
│ ├── adev-dashboard/
│ ├── consumer-saas/
│ └── maine-scientific-compliance/
├── code-samples/ # Sanitized code examples
│ ├── claude-agent-sdk/
│ ├── fastapi-patterns/
│ └── automation/
├── marketing/ # Outreach resources & templates
│ ├── LINKEDIN_UPDATE.md
│ ├── COMMUNITY_ENGAGEMENT_SYSTEM.md
│ ├── READY_TO_POST.md
│ └── GITHUB_PORTFOLIO_STRATEGY.md
├── tutorials/ # Technical guides
└── assets/ # Diagrams, screenshots
Resume and CV available upon request
Interactive Presentations:
- SDK & MCP Workflows Portfolio - Technical deep dive: Claude Agent SDK examples, MCP server development, workflow automation (with ASCII architecture diagrams)
- Terminal Portfolio - Hacker-style terminal interface showcasing the journey from electrical systems to AI systems
- ComEd Energy Efficiency Program - Certified
- OSHA Safety Standards - Certified
Beyond writing code, I deliver measurable business outcomes:
| Metric | Achievement |
|---|---|
| Efficiency Gains | Reduced grant research from 40 hours → 15 minutes (160x improvement) |
| Revenue Generation | $500/month recurring revenue from production systems |
| Process Automation | Eliminated 100+ hours/month of manual workflows |
| Client ROI | 10x cost savings vs manual operations |
| System Reliability | 99.8% uptime on production deployments |
| Grant Discovery | Increased program identification from 3-5 → 20+ per query |
Built production systems coordinating multiple specialized AI agents:
- 15+ Specialized Subagents: Grant research, compliance verification, form filling, agent development, workflow automation, documentation, autonomous execution, session bookmarking, version management, streak tracking, roadmap coordination
- Intelligent Routing: Claude-powered task analysis and agent selection
- Real-time Updates: WebSocket streaming for live user feedback
- Auto Claude Integration: Automated workflows with n8n, Slack webhooks, and scheduled tasks
Beyond client work, I actively explore cutting-edge AI technologies:
- OpenVINO Development - Intel's open-source AI inference optimization toolkit for edge deployment and model optimization
- VCV Rack + AI Integration - Exploring AI-powered modular synthesis and generative audio using VCV Rack virtual Eurorack platform
- MCP Server Ecosystem - Building custom Model Context Protocol servers for specialized workflows (PowerShell automation, browser control, grant tracking)
- Local AI Models - Experimenting with Ollama, local LLM deployment, and privacy-first AI solutions
- Agent Framework R&D - Continuous exploration of multi-agent patterns, tool orchestration, and autonomous systems
Explore sanitized code examples demonstrating production patterns:
- Multi-Agent Orchestration - Pattern for coordinating specialized subagents
- WebSocket Streaming - Real-time agent response streaming
- MCP Server Integration - Custom Model Context Protocol servers
- N8N Workflow Templates - Automation workflow patterns
Blue-Collar Background Meets Cutting-Edge AI
My unconventional path is my competitive advantage. I've operated heavy machinery, conducted energy audits, managed FDA-compliant manufacturing, and coordinated live events. At every job, I identified operational inefficiencies and envisioned technology solutions.
Now I build them.
I understand how businesses actually operate—the permits, the compliance requirements, the stakeholders who need solutions explained without jargon. Combined with deep proficiency in Claude Agent SDK and multi-agent orchestration, I deliver AI solutions that work on day one because they're built by someone who's lived the problems they solve.
- Location: Chicago, IL 60647
- LinkedIn: Connect with me
- GitHub: You're here!
The code samples in this repository are sanitized, generic implementations that demonstrate technical patterns and skills without exposing proprietary client code or intellectual property. They represent the types of systems I build, not exact implementations of client projects.
For live demonstrations or more detailed technical discussions, please contact me directly.
Built with Claude Agent SDK • Deployed on Railway • Powered by Coffee
Available for contract work, consulting, and full-time opportunities