AI systems architect · Automation engineer · BMI2 operator
Work with Tarik Riley.
I bring more than 20 years of experience in IT, telecommunications, systems integration, and automation. My career has focused on making complex systems easier to deploy, test, troubleshoot, and operate.
Today, I apply that experience to building domain-specific agentic AI services and educating others on the engineering, governance, and security practices needed to operate them responsibly.
Round Rock, Texas · Staff and consulting experience across telecom and enterprise systems
Experience
Engineering in real operating environments.
Built integration and regression testing pipelines that transformed a three-week sequential testing cycle into continuous execution across 156 test cases. Earlier automation reduced circuit and order-repair handling from ten hours to one hour per case. Supported cloud infrastructure and security testing for an AWS-hosted IMS/RCS messaging environment.
At Spirent, evaluated SmartNIC and virtualized network platforms through functional, performance, and load testing. Consulted with Verizon engineering teams on 4G/5G automation, development and testing workflows, and process improvements.
Designed infrastructure automation to provision 130 blade servers across two data centers, covering virtualization, Linux, storage, identity, and certificates. Delivered a planned 14-month deployment in 11 months.
Led VMware-to-Kubernetes migration planning and automation on Wind River StarlingX, including GitOps deployment workflows, network provisioning, monitoring, and access controls.
Led Python test automation for a Forescout RADIUS plugin, delivering 110 functional test cases. Introduced structured AI-assisted development and review workflows with validation checkpoints throughout delivery.
BMI2 projects
Building and operating governed AI.
I bring that experience into TestPulse, an evidence-driven diagnostic platform; Fiduciary, a governed AI workflow prototype for capital-investment operations; and A&R, an emerging application of governed workflows to artist development and music-industry operations.
These projects apply a common discipline to different domains: establish the evidence, define authority, evaluate behavior, and preserve human accountability.
Governance, security, observability, and monitoring
An AI service needs clear boundaries around the information it can access, the actions it can propose or execute, and the decisions that require human approval. Evidence provenance, connector health, failures, evaluation results, and audit records must be visible enough to support investigation and informed decisions.
That need also drives AIM—Agentic Integrated Management—BMI2’s work on managing and monitoring AI applications. AIM applies fault, configuration, accounting, performance, and security disciplines to AI operations, alongside business and risk assurance. The goal is to make operational health and governance evidence inspectable throughout an application’s lifecycle.
Contract capabilities
Where I can help your team.
AI architecture & integration
Domain-specific workflows, retrieval and evidence pipelines, APIs, agent tools, and human approval boundaries.
AI SDLC & quality engineering
Requirements traceability, builder/reviewer workflows, automated testing, evaluation harnesses, and release criteria.
Governance & operational visibility
Access controls, audit evidence, failure visibility, monitoring, and operational dashboards.
Platform & infrastructure automation
Kubernetes, GitOps, server provisioning, configuration management, and telecom systems integration.
Learn with me
From business need to accepted change.
Through my LinkedIn series, I document the creation of AI-assisted SDLC pipelines using TestPulse and Fiduciary as practical examples. The series follows planning, requirements, architecture, development, application testing, user acceptance, reporting, and the definition of done—with governance and security carried through the lifecycle.
I serve as the human subject-matter expert, architect, operator, and orchestrator, defining business needs, engineering constraints, and acceptance criteria. AI supports implementation and separate review. Executable tests, evaluation harnesses, and traceable evidence support acceptance decisions.
My teaching extends beyond connecting workflow steps in tools such as n8n. Domain-specific agentic services require deliberate decisions about evidence access, permissions, failure handling, monitoring, and human approval. I share practical lessons to help others build AI services they can understand, govern, secure, and operate.
Read my builder/reviewer SDLC post on LinkedIn ↗Let’s work together
Tell me what you need to build.
I welcome contract inquiries for AI application architecture, AI-assisted SDLC pipelines, test automation, platform integration, and infrastructure automation.
Include your project goals, current environment, expected deliverables, and timeline.
Discuss a contract project ↗support@bmi2.com