Thirty years across engineering, architecture and technology leadership, currently leading a 26-engineer global organization while running Clarity Technologies, the consultancy he founded in 2014 and still operates today.
A conventional résumé would flatten that into a list of roles. The interesting part is the structure: how leadership, delivery, AI work and a pair of junior hockey organizations connect to each other. That is why this site is a graph rather than a timeline.
Keep scrolling and the graph keeps pace with the story. Everything it highlights is readable in full on this page, with or without it.
personroot
Michael Lopez
The person behind the graph. Based in Mesa, Arizona, married with five kids and a house full of cats and dogs. Thirty years across engineering, architecture and technology leadership; when the laptop closes, it's usually a rink, a field, or a flight to Maui.
Mesa, Arizona
Married, five kids
Cats and dogs, in quantity
Travels every chance he gets; Maui is the one he keeps going back to
Two tracks have run in parallel since 2014: leading engineering organizations — individual contributor to Director to VP, global and offshore teams, regulated-industry delivery — and building software independently through Clarity Technologies.
The career is measured in scope and capability rather than a list of logos: financial services since 2004, healthcare and IoT engagements through the consultancy, and an AI and automation practice that stands up the infrastructure and finds the workflows where automation actually pays for itself.
careercareer
Career
A career measured in scope and capability rather than a list of logos. Two tracks running in parallel since 2014: leading engineering organizations, and building software independently through Clarity Technologies.
Founded 2014 and still running. Engineering leadership, architecture strategy and full-cycle delivery for clients across financial services, healthcare and IoT, increasingly centred on AI and agentic automation.
Building teams and the systems of work around them: engineering standards, delivery cadence, architecture review, and the unglamorous discipline that makes releases boring.
Currently leads a 26-engineer global organization
Previously ran 11-person orgs including offshore teams
Leading adoption of AI inside engineering organizations: evaluating models, standing up infrastructure, and finding the workflows where automation actually pays for itself.
Two decades inside lending technology: loan origination, broker systems, point of sale, and the integration surface between them. Regulated, audited, and unforgiving of hand-waving.
Systems that shipped and stayed shipped. Client engagements are described without naming the client; independent work is named directly.
A .NET loan origination platform for one of the largest mortgage brokerages in the United States. An enterprise promotions platform behind more than $50M in global rebate and loyalty programs. Sensor telemetry over LoRaWAN with escalation and auto-resolution. Two commercial point-of-sale products replaced with custom in-house builds — one to a production-ready prototype in ten months.
And the current independent work: a private two-node AI inference cluster running open-weight models locally, a multi-agent orchestration platform with broker-controlled delegation, a healthcare automation system supporting caregivers of dementia patients, three platforms built to run his own hockey organizations, and SafeTube — a parent-controlled viewing platform built by a father of five.
projectprojects
Projects
Systems that shipped and stayed shipped. Client engagements are described without naming the client; independent work is named directly.
A private, two-node inference cluster running open-weight models locally. No data leaving the building. Models are tiered by job: a fast model for routine execution, a reasoning model for analysis, and a burst model brought online only for genuinely hard problems.
Dual GB10 nodes, tensor-parallel over ConnectX-7
Three-tier model routing by task type
Ray-coordinated distributed serving
Allowlisted control plane, no arbitrary shell access
Deterministic tools for arithmetic; models interpret, never calculate
A multi-agent platform where specialised agents are packaged, permissioned and delegated to. A broker ranks candidate agents by capability, permission coverage and queue depth rather than letting agents spawn each other freely.
Broker-controlled delegation, not open peer spawning
Capability packs, skills, rules and hooks
MCP module integration
CRDT-backed collaboration and presence
Role hierarchy with scoped permissions and audit logging
An automation system and companion iOS app supporting caregivers of dementia patients, applying AI-driven decision support and behavioural recommendations at the point of care.
Two separate commercial point-of-sale products replaced with custom in-house platforms. One reached a production-ready prototype in ten months with a hybrid onshore/offshore team.
Two commercial platforms replaced with in-house builds
A .NET loan origination platform serving one of the largest mortgage brokerages in the United States. Engineering strategy, enterprise API architecture, modernization and AI initiatives.
Enterprise APIs across LOS, CRM, ERP and third parties
Sensor and gateway integration over LoRaWAN with event-driven monitoring: configurable thresholds, escalation, automatic resolution, connectivity and battery health, and SMS/email alerting for operational and compliance use.
LoRaWAN and The Things Network integration
Real-time telemetry ingestion and cloud connectivity
An enterprise promotions and incentive platform supporting more than $50M in global rebate and loyalty programs, including a channel incentive engine and over one hundred consumer-facing rebate sites.
A hockey league management platform: registration, rosters, trades and waivers, scheduling, document signing, compliance, finance and statistics. Built as a modular monolith so fifteen business domains stay genuinely separable. In active development, operating his own organizations.
Modular monolith, DDD and clean architecture
15 business modules, each with its own domain layer
ASP.NET Core / .NET 10 with React 19 and TypeScript
The multi-tenant layer above a single team: organizations, subscriptions, usage metering and invoicing, with strict tenant isolation enforced at the database. In active development.
Multi-tenant with composite organization ownership
Junior hockey players live with host families. This matches athletes to host homes: availability windows, compatibility, placement staging and onboarding. An unglamorous, genuinely hard operational problem. In active development.
A parent-controlled viewing platform: every video is reviewed before a child can watch it. AI-assisted review reads metadata and captions, parents approve or block, and revoking access interrupts playback already in progress. Built by a father of five. In active development.
Capability shown through what it built, not through a percentage bar. Every branch here connects to work that evidences it.
AI and agentic systems: agent architecture, retrieval, model routing across providers and local inference, and the operational discipline to run it privately. The .NET/C# through-line of the whole career, down to SDK-level platform extension. Cloud and DevOps discipline that turns deployment from an event into a routine. Data platforms from SQL Server to Qdrant, APIs and integration between systems that were never designed to meet, long-range low-power sensor networks, and deep platform-level knowledge of the mortgage systems the industry actually runs on.
expertiseexpertise
Expertise
Capability shown through what it built, not through a percentage bar. Every branch here connects to the work that evidences it.
Designing systems where models do the judgement and deterministic tools do the arithmetic. Agent architecture, retrieval, orchestration and the operational discipline to run it privately.
Agent architecture and orchestration
Retrieval-augmented generation
Model routing across providers and local inference
Designing how large systems fit together: integration frameworks, API strategy, identity, data movement, and the migration paths that get legacy platforms to somewhere maintainable.
Milestones with a number attached, and the credentials behind them.
A 26-engineer global organization spanning development, automation, platform modernization and AI initiatives. Technology leadership for global rebate and loyalty programs exceeding $50M across hundreds of client brands. A custom point-of-sale platform taken from nothing to a production-ready prototype in ten months. Three actively maintained ICE Mortgage Technology certifications. And a B.A.S. in Technology Management from Northern Arizona University.
achievementachievements
Achievements
Milestones with a number attached, and the credentials behind them.
A custom point-of-sale platform taken from nothing to a production-ready prototype in ten months, replacing a commercial product, with a hybrid onshore/offshore team.
The parts that aren't on the résumé but explain a fair amount of it.
Seven years in Phoenix youth hockey, mentoring young athletes and coaching the same things that matter at work: dedication, teamwork, showing up. Then the step most people don't take — owner/operator of two junior hockey organizations, the Salt Lake Ghost Riders and the Hudson Havoc. Entrepreneurship as a habit: a consultancy since 2014, sports organizations, independent products. And when the laptop closes, football, and a flight to Maui — the one he keeps going back to.
interestinterests
Interests
The parts that aren't on the résumé but explain a fair amount of it.
Hockey, building things, and building organizations
Seven years in youth hockey in Phoenix, Arizona. Mentoring and supporting young athletes, and coaching the same things that matter at work: dedication, teamwork, showing up.
Step back and the graph surfaces two patterns no single line of a résumé states.
Build, don't buy: when a commercial platform stopped fitting the business, it got replaced with a custom build — twice, in two different organizations, one reaching a production-ready prototype in ten months. Analysis still happens first, but when the right tool doesn't exist, build it.
Run it, then build it: he owns and operates two junior hockey organizations and writes the platforms that run them — League Manager, Team Manager, Billet Manager. His own first customer; requirements arrive from the rink rather than from a backlog.
patternpattern.build
Build, don't buy
A recurring decision, surfaced by the graph rather than stated anywhere: when a commercial platform stopped fitting the business, it got replaced with a custom build. Twice, in two different organizations. Analysis still happens first, but when the right tool for the job doesn't already exist, build it. With AI in the mix, that calculus has only shifted further.
Two commercial platforms replaced with custom builds
One reached production-ready prototype in 10 months
He owns and operates two junior hockey organizations, and writes the platforms that run them. League Manager, Team Manager and Billet Manager exist because he needed them himself. He is his own first customer, which means the requirements arrive from the rink rather than from a backlog.
Two junior hockey organizations, owner / operator
Three platforms built to run them
Requirements come from operating the business
An interest became an enterprise became a product line
If the structure above maps to something you're building — consulting, technology leadership, AI and automation — start a conversation. This site deliberately carries no downloadable résumé; the conversation is the path to detail.
Or keep exploring: the full graph is one click away, and every node on this page is already readable above.