I'm an IoT and cloud solutions architect with 35+ years in enterprise IT, including 14 years as enterprise solutions architect at a manufacturing company and co-founder/CTO of an IoT startup where I built end-to-end connected systems from embedded firmware through AWS cloud architecture. My recent portfolio work extends that background into agentic infrastructure operations, AI-assisted DevOps, and applied ML for telemetry-driven monitoring.
- Best fit: IoT Solutions Architect · Cloud Solutions Architect · Platform/DevOps Architect
- Core signal: Secure device-to-cloud systems, agentic AI for infrastructure, Terraform IaC, observability, operator tooling, applied AI infrastructure monitoring
- Recent proof: AWS IoT cold-chain reference stack + autonomous fleet ops AI agent + PyTorch GPU-rack anomaly detection + Go fleet-ops CLI + AI Terraform plan reviewer
End-to-end AWS IoT implementation showing X.509 device identity, MQTT/TLS ingestion, Lambda telemetry processing, DynamoDB storage, CloudWatch observability, and excursion alerting — deployed and validated against real AWS infrastructure.
Why it matters: Demonstrates the same architecture pattern used in real connected-product deployments — not a tutorial, not a diagram. The design decisions section explains the reasoning behind every major choice.
| Project | What it proves | Stack |
|---|---|---|
| iot-ops-agent | Autonomous AI agent for IoT fleet ops — watchdog, incident response, and briefing modes with bounded tool use, structured reasoning logs, and escalation as a first-class outcome | Python · Anthropic Claude · AWS IoT Core · DynamoDB · CloudWatch |
| gpu-rack-anomaly-detection-pytorch | Applied AI infrastructure demo — PyTorch autoencoder trained on simulated GPU rack telemetry, with thermal-derived features and structured anomaly reports for operations handoff | Python · PyTorch · ML engineering · Telemetry · Anomaly detection |
| k8s-inference-ops | Kubernetes deployment-pattern demo for a containerized AI-style inference API — validated locally with kind, two replicas, ConfigMap runtime config, health probes, resource limits, ClusterIP service, and port-forwarded smoke testing | Kubernetes · Docker · Python · kind · Observability |
| iotctl | Operator CLI for IoT fleets — real AWS API calls, no mock data | Go · AWS SDK v2 · Cobra |
| tf-plan-ai-reviewer | AI-assisted IaC risk review with PASS/WARN/BLOCK PR verdicts | Python · Anthropic Claude · OpenAI · GitHub Actions |
| multicloud-sa-toolkit | Multi-cloud Terraform architecture patterns across AWS, Azure, GCP | Terraform · GitHub Actions · AWS · Azure · GCP |
| multicloud-estate-briefing | LLM-generated estate summaries from cloud inventory artifacts | Python · Anthropic Claude · OpenAI · GitHub Actions |
| tf-scaffold-ai | Generates working Terraform scaffolds from plain-language architecture descriptions — constrained generation with security guardrails baked in | Python · OpenAI · Azure OpenAI · GitHub Actions |
- IoT architecture — X.509 device identity, MQTT at scale, cold chain monitoring, edge-to-cloud pipelines
- Agentic AI for infrastructure — bounded tool use, structured reasoning logs, escalation design, responsible deployment
- Applied AI infrastructure — PyTorch anomaly detection, telemetry modeling, structured inference outputs, and operations handoff
- AWS cloud architecture — IoT Core, Lambda, DynamoDB, CloudWatch, IAM — designed and deployed, not just documented
- Infrastructure as Code — Terraform, GitHub Actions, OIDC-based cloud auth
- Operator tooling — CLI design, Go, AWS SDK integration, fleet operations
- AI-enhanced DevOps — LLM review workflows that add signal without replacing engineering judgement
| Domain | Tools |
|---|---|
| IoT | AWS IoT Core · MQTT · X.509 · paho-mqtt · Device Shadow · AWS IoT Jobs · SQS DLQ |
| Cloud | AWS · Azure · GCP |
| IaC | Terraform · GitHub Actions · OIDC keyless auth |
| Languages | Python · Go · C++ (embedded) · HCL |
| Databases | DynamoDB · CloudWatch Logs Insights |
| Observability | CloudWatch · structured JSONL logging · alarm design · runbook execution |
| AI/ML/LLM | PyTorch · anomaly detection · Anthropic Claude · OpenAI · Azure OpenAI · agentic tool use · structured reasoning |
Open to IoT Solutions Architect, Cloud Solutions Architect, Platform/DevOps Architect, and applied AI infrastructure roles. LinkedIn