Agentic system design
Design single- and multi-agent systems that plan, use tools, delegate work, retain context, and escalate when human judgment is required.
Applied AI · Agentic systems · Workflow integration
I turn AI from a chat window into a working part of your business.
I orchestrate specialized AI agents, automate business workflows, combine deterministic testing with AI judgment, and ship AI into production.
I design and deploy agentic workflows that connect models, tools, data, and people—grounded in 18+ years of building dependable production systems. AI makes execution faster; experience provides the judgment to know what right looks like—what to trust, test, reject, and ship.
- Multi-agent orchestration
- Business workflow automation
- Human-in-the-loop systems
applied-ai.js
const appliedAI = {
experience: '18+ years',
builds: ['agents', 'workflows'],
connects: ['models', 'tools', 'people'],
tested: true,
observable: true,
humanApproved: true,
shipsToProduction() {
return (
this.tested &&
this.observable &&
this.humanApproved
);
}
};AI expertise
Beyond prompting. I build the system around the model.
Useful AI is an engineering and operations problem. I combine model capabilities with software architecture, business context, and responsible controls to move AI from experimentation into dependable day-to-day work.
Workflow integration
Connect AI to the systems where work already happens—APIs, internal tools, content platforms, ticketing, data stores, and operational processes.
AI-enabled delivery
Equip engineering and business teams with repeatable agent workflows for research, implementation, review, documentation, and quality assurance.
Production guardrails
Build for observability, evaluation, permissions, cost control, data boundaries, and human approval—not just a convincing prototype.
From opportunity to operating capability
- 01
Find leverage
Identify high-value work where AI can improve speed, quality, or capacity.
- 02
Design the system
Choose the right model, tools, data access, agent boundaries, and human checkpoints.
- 03
Integrate the workflow
Connect the system to real software, teams, and operating processes.
- 04
Evaluate with experience
Recognize correctness, test reliability and business impact, and refine the system over time.
human_judgment
Experience recognizes correctness.
AI can generate an answer that looks plausible. Knowing what right looks like—seeing whether it fits the real system, catching what is subtly wrong, and deciding what is safe and ready to ship—is earned through experience.
Selected impact
AI expertise built on years of shipping consequential systems.
A representative view of the work behind the title—from AI-enabled engineering operations to high-scale logistics and enterprise digital platforms.
Applied AI · Accessibility engineering
AI-assisted accessibility assurance
Designed a reusable Zivtech harness that combines deterministic automated testing with nondeterministic agent review. Applied on the NCLC platform, it covers both machine-testable failures and higher-context issues involving semantics, keyboard behavior, state communication, and user experience.
Moves accessibility earlier into delivery, expands coverage beyond conventional scanners, and keeps experienced human judgment in the approval loop.
AI adoption · Operating model
Zivtech’s company-wide AI-first practice
Driving a company-wide mandate to make AI a standard part of how Zivtech operates—not a side tool. That includes agent-enabled research, specification, engineering, code review, documentation, quality assurance, accessibility, and internal workflows, with clear human accountability.
Turning isolated experimentation into repeatable workflows, shared standards, and production integrations across the company.
Platform engineering · Logistics
GroScale
Served as lead architect and developer for a high-throughput shipping platform integrating NetSuite, multiple carriers, label generation, tracking, and exception handling through APIs and webhooks.
Scaled from 10,000 to 500,000+ packages per month—a 50× increase—while handling peak loads without downtime.
Architecture · Publishing and commerce
NCLC Digital Library
Led the modernization of a subscription publishing platform from Drupal 7 to Drupal 10, including a React book reader, Solr search, automated content imports, subscription commerce, and the AI-assisted accessibility workflow.
A production-ready platform that streamlined editorial operations and improved access to authoritative legal resources.
Architecture · Complex digital portfolios
Enterprise and multi-brand platforms
Delivered architecture and engineering across institutional, enterprise, and consumer-brand environments, including Children’s Hospital of Philadelphia, Johns Hopkins, Merck, LSAC, and Philly Marketing Labs.
Production digital systems with demanding content operations, integrations, governance, security, and accessibility requirements.
Current community project
Practical technology for organizations of every size.
Enterprise experience, applied with the speed and resourcefulness small organizations need.
Augusta, Georgia
Backyard Bullies Wrestling Club
Supporting a growing community wrestling organization across research, information architecture, UX and visual design, content strategy, and web implementation. The result gives families one clear place to understand programs, schedules, camps, coaches, membership, and ways to get involved.
I'm also helping the club grow its funding base—assisting with sponsorship recruitment research and grant research to identify and pursue opportunities that sustain its programs.
This is the same approach I bring to small-business engagements: use AI to accelerate research and execution, then apply experienced human judgment to the strategy, design, and finished product.
Visit Backyard Bullies Wrestling Club(opens in new window)About me
Turning complex technology into systems people rely on.
I'm Jonathan DeLaigle—an applied AI architect, CTO, and software engineer with 18+ years of experience turning complex technology into dependable business systems. At Zivtech, I lead a company-wide AI-first mandate that moves AI beyond individual tools and into repeatable engineering, quality, accessibility, documentation, and business workflows. I design agentic systems that can reason across tasks, use tools, collaborate, and connect to real operational processes, grounded in deep experience delivering for organizations including Children’s Hospital of Philadelphia, Johns Hopkins, Merck, LSAC, GroScale, and NCLC.
While I currently serve as CTO at Zivtech, I take on select advisory, consulting, and community-focused engagements. That includes AI workflow strategy and agentic systems as well as practical research, product design, web delivery, and early-stage technical direction for small businesses and growing organizations.
Experience
(Aug 2020 - Present)
Chief Technology Officer
Zivtech
(Aug 2016 - 2020)
Senior Software Engineer & Co-Founder
GollyGood Software
(May 2014 - Aug 2016)
Senior Software Engineer
Independent Consultant
(July 2013 - May 2014)
Senior Drupal Developer
Mediacurrent
(Aug 2008 - Dec 2013)
Web Developer
Advomatic
(Jul 2006 - Aug 2008)
Associate Web Developer
Morris DigitalWorks
Technical Foundation
AI systems only create value when the surrounding software is sound. My work is grounded in production engineering, cloud infrastructure, automation, APIs, and technical delivery.
Education
(Aug 2007 - May 2018)
Bachelor of Science in Computer Science, Certificate in Cyber Defender; GPA: Cum Laude
Augusta University
(Aug 2004 - May 2007)
Associate of Applied Science in Computer Programming
Augusta Technical College
Contact
Let's put AI to work in your business.
Have an AI workflow to design, a product to shape, or a small-business idea that needs experienced technical execution? Tell me what you’re trying to accomplish.