Most AI pilots die between the demo and your actual data.
I embed with your team and ship agents that run against production systems — the identity plumbing, the retrieval layer, the evals, the failure modes. Fifteen years of IT and a decade of cloud and DevOps sit behind every one of them.
What I actually do
Three things, and they compound. The agent work is only credible because the platform work is real.
Forward-deployed AI engineering
I work inside your environment, not next to it. Discovery, build, and hand-off — agents wired into your data, your identity model, and your release process.
Agentic & retrieval systems
Knowledge-graph retrieval and GraphRAG, natural-language-to-SQL, model-as-tool orchestration, and the evaluation harnesses that tell you whether any of it works.
Cloud platform & reliability
Azure and AWS, Kubernetes, Terraform and Bicep, CI/CD, disaster recovery. The unglamorous layer that decides whether the AI stays up.
Fifteen years before the agents
A decade of building and running the cloud, data and ML platforms these systems sit on top of. That's why the plumbing questions — identity, permissions, what happens when a dependency quietly returns nothing — are familiar rather than theoretical.
-
2026 – PRESENT · MICROSOFT (CONTRACT)
Cloud AI Engineer — agents in production
Lead engineer on a knowledge-retrieval agent workstream for a global materials manufacturer: prompt and knowledge agents on Foundry Agent Service, a 120 GB Azure AI Search knowledge architecture, model-as-tool integration over OpenAPI/MCP, and a 374,000-node ontology queried through GQL and lakehouse SQL. Separately, Data & AI Engineer lead on a 16-week MVP for a global beverage manufacturer, delivered in the customer's own tenant with a 29-person joint team.
-
2025 – 2026 · CISCO (CONTRACT)
Cloud Engineer — FedRAMP High
Provisioned and onboarded applications inside a FedRAMP High environment using Terraform and Argo CD GitOps, holding every infrastructure change to federal security standards. Architected HashiCorp Vault Secrets Operator with automated credential rotation.
-
2024 – 2025 · AMC (CONTRACT)
Cloud/DevOps Engineer — ML and data platform
Bicep modules provisioning Azure ML workspaces, Synapse Analytics and Data Factory — the platform layer ML actually runs on. Zero-downtime AKS upgrades, multi-environment GitHub Actions, observability wired into the deploy pipeline.
-
2022 – 2024 · THE STANDARD
Cloud/DevOps Engineer — multi-region at scale
Multi-region Azure architecture in Bicep and Terraform, AWS infrastructure in TypeScript (Pulumi/CDK) across EKS, VPC and RDS, Azure AD least-privilege IAM, VM and network hardening to pass compliance audits, and an offshore engineering team to deliver it.
MS Systems Engineering & BS Electrical Engineering, California State University, Fullerton. Certified: Azure AI Engineer (AI-102), AI-900, DP-900, AZ-900, Terraform Associate, KCNA. Capability statement →
Work with me
Remote, US-based. Available for contract and subcontract engagements, including federal and defense work.