I'm a senior full stack developer with 15+ years of enterprise experience who has pivoted into building production AI tooling. I don't just study AI -- I ship it. After earning 12 AI certifications, I immediately applied that knowledge to build production tools: a local-first AI coding assistant, a multi-agent PR review system, a legacy migration toolkit, and a structured debugging framework for AI coding agents. My approach combines deep enterprise experience with hands-on AI engineering and collaborative problem-solving.
- Design-First Mindset: I believe great software starts with great design. Every solution I build prioritizes user experience and clean architecture.
- Strategic Problem Solver: I approach challenges methodically, breaking down complex problems into elegant, maintainable solutions.
- Team-Focused Leadership: My success is measured by my team's success. I foster collaboration, knowledge sharing, and win-win outcomes.
- Learn, Build, Ship: I don't stop at certifications. I completed 12 AI courses and immediately built production tools with that knowledge.
- Web Services: RESTful APIs, SOAP
- Message Brokers: ActiveMQ, RabbitMQ
- Protocols & APIs: LDAP, JDBC, JNDI, AJAX
- ✨ Enterprise-grade architecture and clean code principles
- 🎨 Design-driven development with user-centric focus
- 🤝 Agile collaboration and pair programming
- 📊 Data-driven decision making and performance optimization
- 🔄 CI/CD pipelines and continuous improvement
I identify real developer pain points, build tools that fix them, and put them in developers' hands. All three VS Code extensions are live on the marketplace -- rated ⭐⭐⭐⭐⭐ by early users.
Every AI coding tool on the market has the same fine print: your code leaves your machine. I found that unacceptable for enterprise teams and security-conscious developers -- so I built the alternative. Safe Agent is a privacy-first AI coding assistant that runs 100% on your machine. CLI, HTTP service, and VS Code extension -- all talking to your local models. No API keys. No telemetry. No surprises. Why developers choose it:
|
Every developer wants a great AI coding assistant. Not every developer wants a $50/month bill to go with it. I built ZenCoder to solve exactly that -- intelligent AI assistance that routes to free and near-free models first, so the quality stays high and the cost stays near zero. ZenCoder is a smart-routing AI coding assistant that dispatches each request to the best available model automatically -- local for simple tasks, cloud for complex ones. Bring your own keys for 7 providers, or run entirely on local models. Why developers choose it:
|
I watched AI coding assistants make the same mistake over and over: read an error, assume the cause, generate 10 wrong fixes, burn thousands of tokens, and loop. The problem isn't the model -- it's that the AI never asks. I built AgentProbe to fix that. AgentProbe gives GitHub Copilot 25 MCP tools for structured, Socratic debugging. It asks one targeted question before touching a single line of code -- and resolves 80%+ of common errors locally with zero AI cost, using a built-in rule engine of 30+ JS/TS/Java patterns. The three laws AgentProbe enforces:
Why it matters:
|
|
JBoss-to-Spring Boot migrations fail because AI tools try to swallow an entire legacy codebase in one prompt -- and lose critical code in the process. I've done these migrations for 14 years. I knew exactly where they break. So I built the toolkit that prevents it. A Python-based automation framework that runs a 5-phase migration workflow -- Analyze, Plan, Generate Prompts, Detect Code Loss, Repair -- designed to work with any AI coding agent (Copilot, Cursor, Claude Code). What makes it different:
|
|
I completed 12 AI certifications and immediately applied every concept to build real tools. The table below maps each certification to the tangible outcome it produced.
| Certification | Provider | What I Built With It |
|---|---|---|
| Introduction to Model Context Protocol | Anthropic | Context management system in Safe Agent's VSCode extension |
| Claude Code in Action | Anthropic Education | Safe Agent's interactive CLI shell and git automation workflow |
| Advanced Prompt Engineering for Everyone | Vanderbilt University | Prompt design for CrewAI's review agents; chunked prompt strategy in Legacy Migration Toolkit |
| Introduction to Generative AI for Software Development | DeepLearning.AI | Multi-LLM provider architecture in Safe Agent (Ollama + LM Studio) |
| Agentic AI and AI Agents: A Primer for Leaders | Vanderbilt University | CrewAI PR Workflow's multi-agent orchestration and task decomposition |
| Software Design and Architecture - Object-Oriented Design | University of Alberta | SOLID principle gate in CrewAI PR Workflow; clean architecture across all three tools |
| AI-Powered Software and System Design | DeepLearning.AI | 5-phase migration workflow architecture in Legacy Migration Toolkit |
| AI for Software Engineers | Coursera | AI-agent-agnostic design in Legacy Migration Toolkit -- works with any coding agent |
| Team Software Engineering with AI | DeepLearning.AI | Collaborative AI workflows in CrewAI PR Workflow's multi-agent review pipeline |
| Building Applications with DeepSeek | Board Infinity | DeepSeek Reasoner integration for performance and security analysis in CrewAI PR Workflow |
| Deploy AI Apps with Cloudflare | Scrimba | Cloud deployment patterns for AI-powered services |
| AI Governance and Technology Foresight | Coursera | Responsible AI design decisions across all tools; governance-aware architecture in AgentProbe |
Every certification I earn gets stress-tested in a real project. Learning without building is just theory.
- 🔧 Building AI developer tools: Actively shipping Safe Agent, ZenCoder AI, Legacy Migration Toolkit, and AgentProbe -- all live on the VS Code Marketplace and PyPI
- 🧠 Multi-agent systems: Designing agent pipelines that decompose complex tasks (code review, migration analysis) into specialized, sequential AI passes
- 🐛 AI agent debuggability: Building AgentProbe to give AI coding agents structured, actionable diagnostics instead of cryptic raw errors
- 🔁 Resilient AI workflows: Solving real-world friction like provider rate limits to keep multi-step pipelines running without losing state
- 🔒 Privacy-first AI: Championing local LLM architectures that keep code and data off third-party servers
- 🏗️ Enterprise legacy modernization: Combining 14 years of Java/JBoss expertise with AI-driven migration automation
- 🌱 Mentoring teams on AI adoption: Helping teammates integrate AI tools into their workflows with practical, hands-on guidance
When I'm not developing software, you'll find me nurturing a different kind of growth—my own food garden! I'm passionate about sustainable living and deeply connected to nature.
- 🥬 Home Gardener: I grow my own food, practicing sustainable agriculture and understanding the value of patience and nurturing—skills that translate beautifully into software development
- 🌍 Sustainability Advocate: Committed to living in harmony with nature and making environmentally conscious choices
- 🌿 Nature Enthusiast: Finding inspiration and balance through my connection with the natural world
Just as I cultivate plants, I cultivate code—both require care, attention to detail, and a long-term vision for growth.
I'm always interested in collaborating on innovative projects and connecting with fellow developers who value teamwork and continuous growth.
"I don't just study AI—I build with it. I don't just earn certifications—I ship the tools they teach me to create. Success isn't about credentials on a wall; it's about code in production that makes teams better."