john mark lowry

director, product & innovation · applied AI + platform architecture · dallas, tx johnmark.lowry@gmail.com · linkedin.com/in/johnmarklowry · github.com/johnmarklowry · johnmarklowry.com

Product leader and applied AI builder working in the space between product and engineering. Owns the Product function for Team One's Lexus and Toyota digital business and the roadmap for the Lexus.com ecosystem (~7–8M monthly users). Hands-on builder of 20+ production AI tools in use across the agency and of both launch applications for the Toyota GR GT US launch, written through AI-assisted engineering on top of 14+ years of hand-written code.

---

experience

director, product & innovation — Team One (Publicis Groupe)

apr 2022 – present · dallas, tx

product leadership

• Own the agency's Product function end to end: product owner practice, backlog and acceptance-criteria standards, roadmaps and KPIs, platform prioritization, triage operations, and Product hiring and people leadership. Product is the single decision owner across a 98-person digital organization (8 scrum teams)
• Own the product roadmap for the Lexus.com ecosystem (7–8M monthly users). Led the Lexus and Toyota Build & Price rebuild (700K monthly users) as the first agency-side product owner, to JD Power best in class vehicle configurator across all OEMs
• Founded the agency-side product owner discipline and scaled it from 1 role to a team of 6; merged account and product owner roles into a single Product Lead per scrum team, with the coverage model, training plan and ramp to make it stick
• Diagnosed delivery latency across ~10,000 Jira issues and authored a Delivery Lead operating model and charter, moving backlog readiness out of production and into product

Toyota GR GT launch (Gazoo Racing)

• Built the proof of concept for the GR GT in-dealer experience — a real-time Unreal Engine 5 configurator driving an LED video wall from a tablet — outside the production process. Engineering adopted the prototype as the launch application instead of rebuilding it
• Personally implementing both launch applications through AI-assisted development (Claude Code): a web configurator with live cloud pixel streaming, and the in-dealer experience. Enterprise engineering owns integration, infrastructure and QA; those PRs go through the same review as any engineer's
• Moved the launch configurator out of the AEM monolith into a standalone static app that consumes shared component libraries, taking the CMS off the critical path. Defined the build-vs-deploy model with hardware vendors: the agency ships versioned golden builds; vendors image, roll out and run L1/L2 support
• Coordinate three scrum teams and five external partners; ran the configurator demo on site at Monterey Car Week 2026

applied AI and engineering

• Moved AI from experiment to infrastructure: 20+ production tools (plus proofs of concept) on the agency's internal AWS/EKS platform, 100–250 daily active users, several hundred hours saved per month. Migrated 15 apps onto the platform and instrumented 18 with shared LLM cost and usage telemetry
• Built the agentic triage bot (Jira + Slack + LLM): ~90% less manual routing, severity calls in ~5 minutes instead of up to a day, recommendations accepted ~99% of the time
• Diagnose root causes of production defects on Lexus.com and the inventory platform straight from source code and live APIs (pricing logic, video codecs, digital asset management, CMS templates), usually symptom to fix path in a day. Teaching the agent-based root-cause analysis workflow to the engineering org
• Trained the 400-person agency on AI (150–200 in active adoption); author the playbooks and run the AI Summer Sessions series, including "Judgment in the Loop" on verifying AI output

senior experience consultant — Slack Technologies

dec 2021 – apr 2022 · portland, or

• Fortune 500 Slack deployments (10K+ users): collaboration architecture and change management. A deliberate experiment that confirmed a preference for owning systems over advising on them

senior client partner, product owner — Team One (Publicis Groupe)

apr 2020 – dec 2021

• Technology strategy for a $100M+ digital portfolio; selected for the organization's most complex, highest-profile initiatives

client partner, product owner — Team One (Publicis Groupe)

apr 2018 – apr 2020

• Led cross-functional teams delivering $50M+ in annual value; worked directly in AEM and developer consoles to cut translation loops between client and technical teams

associate director, digital products — Agency Creative

may 2016 – apr 2018 · dallas, tx

• Brought outsourced development in-house: project costs down 30–40%, profitability up; digital went from production service to strategic capability

senior art director, front end web developer — United Networks of America

jun 2014 – may 2016 · baton rouge, la

• Marketing systems for prescription benefits programs: PHP, WordPress, marketing automation

art director, print production manager — BizCard Express

feb 2013 – jun 2014

• Layout and CMYK production across signage, vehicle wraps, vinyl and print — where the respect for production constraints was forged

developer, designer, applied AI — freelance

jan 2011 – present

• Predictive models for market and ad-buying analysis (TensorFlow, PyTorch); custom web development and design

---

projects

• Compass, AI work intake and roadmap hub — production in 4 days (Next.js, Postgres, 229 tests). Found the schema already in Jira, cancelled an AI backfill, and cut active-but-unowned work from 25 items to 1.
• Claim Check, AI fact verification — code must prove every quoted source passage exists before a model may judge a claim. Caught all 5 planted errors in a seeded test, at about $1 per run.
• Lens, cumulative analytics — built in a day when API access was declined: any CSV export becomes a persistent, queryable dataset joined to the media warehouse, with scheduled AI briefings tested against real queries.
• Reading the repo as a product owner — root causes for production defects found from source and live APIs, symptom to fix path in about a day; the agent workflow is now taught to the engineering org.

---

skills

product: product management, product strategy, roadmaps and KPIs, platform prioritization, product operating models, cross-functional and stakeholder leadership, people management and hiring, agile/scrum, backlog and acceptance criteria (Gherkin), Jira, Confluence applied AI: generative AI, LLMs, AI agents and agentic workflows, RAG, MCP servers, evals and output verification, prompt engineering, AI-assisted development (Claude Code), Anthropic Claude API, OpenAI API, Google Gemini, machine learning (TensorFlow, PyTorch), AI governance engineering: TypeScript, JavaScript, Node.js, React, Next.js, Python, PostgreSQL, Prisma, Redis, REST APIs, WebSockets, Docker, AWS (EKS, S3), Railway, CI/CD platforms and 3D: Adobe Experience Manager (AEM), vehicle configurators, Unreal Engine 5, Pixel Streaming, real-time 3D, design-token systems method: product spec and acceptance criteria → AI-assisted build (Claude Code) → validation → production ownership

---

education & awards

• Louisiana State University — business administration / marketing (2012–2013)
• Baylor University — music theory & composition / jazz guitar (2007–2011)
• Google UX design professional · advanced analytics · growth-driven design
• JD Power best in class vehicle configurator (Lexus build & price, across all OEMs)
• AI transformation team lead · national mentorship program leader