AI Infrastructure
Enterprise AI reference architecture - (Capital One - Discover)
A governed AI platform pattern for regulated enterprise environments balancing experimentation, evaluation, security, and operations.
Welcome to
The Evolving Stack
While we are here, we must change to grow.
Engineering & Architecture Leader | Hands-On Builder | Technical Strategy & Execution
Systems Engineering
Featured systems
A collection of systems I've ideated, designed or led (or all three).
AI Infrastructure
Local inference runtime for developer productivity (Capital One - Discover)
A privacy-preserving local model runtime pattern for enterprise developer tooling.
Governance
Compliance-as-code architecture (Capital One - Discover)
A policy-as-code pattern for turning control requirements into queryable, testable platform state.
Enterprise Infrastructure
Global virtual desktop platform (Caterpillar, Inc.)
A remote-work platform pattern for secure access, endpoint consistency, and resilient user experience across regions.
Platform Engineering
Platform automation control plane (Capital One - Discover)
An infrastructure automation pattern for repeatable provisioning, lifecycle control, and operational visibility.
Complete index
View all systems
Browse the full public collection of system writeups, including featured and supporting architecture examples.
Architecture patterns
Good technology work should rarely be one-off.
This section exposes some pattern ideas around modern technology architecture. It emphasizes the tradeoffs and operating model behind platform decisions: what gets governed, what gets automated, and what must stay reviewable.
Start with the problem, not the platform
Tool choices come later. I start by understanding the workload, the constraints, the people who have to run it, and what success actually looks like.
Design for the whole lifecycle
Provisioning is the easy part. Good platforms also account for ownership, change, drift, support, failure, and eventually retirement.
Make the system explain itself
The more infrastructure, policy, and platform state can be inspected and queried, the less teams have to rely on tribal knowledge and one-off review.
Put workloads where they belong
Cloud, on-prem, and edge are all valid answers. The right one depends on cost, latency, resilience, data, licensing, security, and the team expected to operate it.
Treat AI like a platform, not a demo
Model access is only one piece. Production AI also needs evaluation, observability, policy boundaries, ownership, and a clear path from experiment to something teams can trust.
Leave something others can use
A good architecture should outlive the meeting that created it. I try to leave behind patterns, decisions, automation, and enough context for the next person to keep moving.
Recognition & Trust
Responsibility
Recognition earned through sustained technical leadership, meaningful enterprise impact, and the trust to take on increasingly difficult work.
Enterprise Recognition
Capital One-Discover Acquisition Integration (2025-2026)
Tasked to lead merger-acquisition decisions for IaC and ConfigMgmt Domain toolings.
Enterprise Recognition
Capital One-Discover Project Showcase (2025)
Selected to showcase local coding assistant project as evidence of State of the Art AI development.
Enterprise Recognition
Discover Domain Lead Nomination (2023)
Nominated as Domain Architect Expert for Infrastructure as Code, Configuration Management and broader Automation Architecture Team.
Enterprise Recognition
Caterpillar CEO Award (2022)
Received Caterpillar's CEO Award in recognition of exceptional contribution and impact.
Performance Recognition
Caterpillar Outstanding Individual Performance (2022)
Recognized for outstanding individual performance and contributions to enterprise technology.
Technology Leadership
Caterpillar CIO Award Nominee (2021)
Nominated for a CIO Award for technology leadership and impact.
Career Growth
Caterpillar Top Performance Rating (2021)
Earned the highest performance rating possible, leading into continued growth and increased responsibility.
Military Recognition
Airman of the Deployment — Saudi Arabia (2000)
Recognized for outstanding performance during deployment while serving as an F-15 Crew Chief and Flight Leader - 27th Fighter Squadron.
Technical themes
Branching out
A little more about the projects, technologies, teams, and experiences that have shaped how I work.
Built and shaped
Enterprise AI reference architectures · Enterprise search and RAG systems · Local AI inference platforms · Global virtual desktop platforms · Regulated VDI environments · Infrastructure automation · Workload placement strategy
Worked across
AWS · Azure · OpenShift · Kubernetes · Terraform · Ansible Automation Platform · Citrix · Intune · SCCM · Crossplane
Worked with
Enterprise Architecture · Infrastructure Architecture · Cloud Platforms · AI Platforms · End User Computing · Automation · Security · Governance & Review · Application Teams
Experience across
Financial Services · Manufacturing · Regulated Technology · Hybrid Cloud · On-Prem Infrastructure · Global Enterprise
Worked around the world
International experience including the 2014 SEM acquisition in China, helping connect acquired operations through ERP segmentation, Citrix-based application access, and better operational visibility.
Served before systems
United States Air Force · F-15 Crew Chief · 94th Fighter Squadron · Deployed to Prince Sultan Air Base, Saudi Arabia
Leadership
Freedom is the core belief.
Freedom is the principle I keep coming back to. Give capable people room to think, decide, and build, then stay close enough to help when the path gets unclear. I prefer regular check-ins over constant oversight, clear direction over control, and trust that is earned in both directions.