I'm a hands-on engineering leader with 25+ years of experience building scalable backend systems, APIs, and cloud-native solutions across Fortune 1000 companies and startups. I specialize in Java, Spring Boot, microservices, and cloud platforms like AWS, GCP, and Azure.
- Building agentic AI workflows with LangChain, Streamlit, and Ollama
- Developing real-time trading dashboards with LSTM models, sentiment feeds, and Alpaca integration
- Creating research assistants using Graph-based RAG (Retrieval-Augmented Generation)
- Running self-hosted LLMs (LLaMA, Mistral) for private inference and automation
- Experimenting with Raspberry Pi IoT projects like smart garage door control
- Exploring agentic AI agents (autonomous workflows, planning & reasoning)
- Prototyping AI-powered local SEO & outreach agents with Python + LangChain
- Enterprise microservices & APIs with Java/Spring Boot
- Full-stack apps with React, TypeScript, and FastAPI
- CI/CD + observability with GitHub Actions, Jenkins, Docker, Kubernetes, and Grafana
- Data systems: BigQuery, Spanner, MongoDB, Neo4j (graph queries)
- AI Trading Bot (LSTM + live dashboard)
- Agentic Research Assistant (Graph-RAG powered)
- Raspberry Pi Smart Garage (local-first control)
- Niche Radar AI (Shopify/AliExpress product research agent)
- Real Estate Deal Screener (ARV prediction + AI insights, Streamlit → React/FastAPI)
- AI-Powered Chat UI (Next.js + Tailwind CSS + Ollama streaming backend)
- AI Consultant Portfolio (Astro + Netlify, with Medium integration)
✍️ Building an AISecOps Runtime: Securing RAG and Agentic AI Systems with Real-Time Telemetry
AI systems are no longer static models behind APIs. They retrieve external data. They call tools. They execute workflows. And that means they can be attacked. While most discussions focus on model sa…
🔐 A Threat Model for Agentic AI (MCP, A2A & Swarm Systems)
There’s a dangerous misunderstanding spreading across the agentic AI ecosystem: Guardrails are not content filters. In single-model chat systems, safety meant blocking harmful outputs. In agentic sys…
✍️ Securing AI Agents in the Enterprise: Building an AISecOps Plugin for OpenClaw
AI agents are moving from experimentation to execution. They create Jira tickets. They deploy infrastructure. They query internal systems. They mutate production data. That’s powerful. It’s also dang…
✍️ Sandboxing AI Tools in OpenClaw: A Practical AISecOps Pattern
From “Cool POC” to Practical AISecOps Architecture Most people experimenting with OpenClaw run everything on a single machine:
Gateway Agent Tool execution API keys
It works. It’s convenient. It’s …
Principles for Securing Agentic Systems As AI systems move from passive chat interfaces to autonomous agents with tool access, traditional DevSecOps is no longer sufficient. AISecOps is the disciplin…
- Built with Raspberry Pi, MQTT, and Siri Shortcuts. Includes remote control UI and door status monitoring.
- Visualizes call history patterns for supervisors (e.g., transfers, holds). Built with Spring Boot, React, and MariaDB.
- Streamed live TV over the internet using a TV tuner card, VLC, and Windows Media Encoder.
- Built before mainstream streaming platforms existed.
- After Napster was blocked, I prototyped a simple decentralized file-sharing tool in Java.
- Shared broadband securely with neighbors using consumer routers and custom firmware — an early look at community networks.
- Created for a grad school project — featured menu browsing and order submission well before Grubhub or DoorDash.
- 🌐 Portfolio
- 🧪 Medium Articles
“Innovation is not just building for today — it's solving problems before people know they have them.”
Thanks for stopping by! ⭐️