IN

AI, after the demo.

Product judgment for systems that learn.

Work
01

Now

Capability is only the beginning.

Most AI demos fail in the distance between what a model can do and what a product can reliably and consistently deliver. That distance is where the real work exists.

01Multi-agent systemsOrchestration that earns its complexity.

Custom orchestration, role design, handoffs, guardrails, and recursive workflows built around accountable outcomes, not agent count.

02Knowledge + memoryContext that compounds instead of disappearing.

Graph-based persistent memory and AI-readable knowledge models that make context durable, queryable, and reusable across products and teams.

03Evals + experimentationEvidence before confidence.

Custom harnesses, real-time KPI visibility, and feedback loops that make quality measurable and give intelligent systems a path to improve.

040-to-1 product leadershipThe whole path from maybe to useful.

Research, business case, product architecture, team alignment, launch, and adoption. Technical enough to see the system; clear enough to align the room.

02

Selected work

Three projects. Four questions.

Business problem. Insight. Solution. Impact.

GoDaddy Research2026A five-year data problem, resolved in production.20M+ records / under 3 days

Business problem

A data problem had remained unresolved for more than five years. The projected fix required two to three quarters and additional headcount.

Insight

The constraint was not raw model capability. It was how the work was decomposed, governed, and evaluated across the entire dataset.

Solution

Isaac designed and deployed a governance-aware multi-agent data orchestration system with explicit roles, controlled handoffs, and full-dataset evaluation.

Impact

The production system shipped in under three days and achieved 100% evaluation accuracy across more than 20 million records.

  • Multi-agent
  • Data orchestration
  • Evals
  • Production AI
Current LinkedIn profile (opens in a new tab)Public description of the five-year problem, production deployment, evaluation result, and projected alternative requiring two to three quarters and additional headcount.
NoFraud2024-2025From monolith to momentum.12x release velocity / ~50% fewer resources

Business problem

A monolithic ML risk engine made product changes slow, resource-intensive, and difficult to test after a company restructure.

Insight

Fraud decisioning needs both repeatability and judgment: deterministic controls where consistency matters, and non-deterministic AI + ML models where signals and uncertainty matter.

Solution

Isaac rebuilt the core product and experimentation infrastructure as a hybrid AI + ML decision system, then redesigned onboarding and the platform taxonomy around it.

Impact

Release velocity increased 12x, resource requirements fell about 50%, onboarding conversion increased 300%, and cost of goods sold fell about 25%.

  • AI + ML
  • Decision systems
  • Experimentation
  • Fintech
Current LinkedIn profile (opens in a new tab)Public role summary and reported product outcomes.
GoDaddy2022-PresentFrom a domain purchase to a business launch.First generative AI product / used by millions

Business problem

First-time founders were buying domains, then stalling at the harder work: choosing a name, building a brand, incorporating, launching a site, and finding customers.

Insight

The domain purchase was not the destination. GoDaddy's proprietary data and emerging generative AI could connect the steps that followed into one coherent product experience.

Solution

Isaac identified the opportunity, drove the business case, and led the early research, ideation, and UX for what became Airo.

Impact

Airo became GoDaddy's first generative AI product and is now used by millions of entrepreneurs to move from an idea to an operating business.

  • Generative AI
  • 0-to-1
  • Product architecture
  • GTM
GoDaddy Airo (opens in a new tab)Official product page and current product scope.
03

Selected outcomes

What changed in production.

Selected outcomes from systems I helped build, rebuild, or bring to market.

20M+

production records evaluated

A production multi-agent data orchestration system achieved 100% evaluation accuracy and shipped in under three days.

12x

release velocity

A hybrid AI + ML decision system combined deterministic controls with non-deterministic models, moving NoFraud from monolith to momentum.

$3M

monthly revenue, built from zero

An e-commerce platform scaled from zero to $3M per month in under one year.

300%

conversion increase

A clearer onboarding experience and platform taxonomy turned structural product work into measurable growth.

04

Timeline

The platforms change. The pattern holds.

Across twenty-five years, the work has stayed consistent: recognize the shift early, build what makes it useful, and make it scale.

  1. AI product systemsIntelligence became an operating layer.

    GoDaddy / Product Manager → Principal PM, AI Initiatives

    Joined as Product Manager in 2020 and became Principal Product Manager, AI Initiatives in 2022. Leads AI-native product development across multi-agent orchestration, graph-based memory, evaluation frameworks, and recursive learning. In 2026, a multi-agent system shipped to production in under three days and achieved 100% evaluation accuracy across more than 20 million records. He also built an org-wide experimentation agent that brings real-time KPI visibility and scientific rigor to non-engineering teams.

    • Multi-agent
    • Knowledge graphs
    • Evals
    • Generative AI
    Current LinkedIn profile (opens in a new tab)Current role, public focus, reported product outcomes, and career chronology. Verified July 2026.
  2. Concurrent platform rebuildChanging the engine while it was running.

    NoFraud / Principal Product Manager (Contract)

    Refactored a monolithic ML risk engine into an AI-driven decision system, cut resource requirements by about half, increased release velocity 12x, and redesigned onboarding to improve conversion 300%.

    • AI decisions
    • Risk
    • Shopify
    • Platform rebuild
  3. The 0-to-1 yearsA product studio for consequential bets.

    Wonderful / Head of Product (Chief Futurist)

    Led new SaaS, fintech, commerce, and consumer products across partner companies, including GoDaddy. Led product for the world's first smart hydration system for NASCAR racers, shipped iOS products recognized as Editor's Picks, and scaled an e-commerce platform from zero to $3M per month. Reworked the taxonomy, plugins, and UX for APM Music's licensing platform, used by Netflix, Apple, and NBCUniversal, and led the rearchitecture and rebuild of Eventim's U.S. live-events ticketing platform (formerly See Tickets).

    • 0-to-1
    • Fintech
    • Platform architecture
    • GoDaddy
  4. TelemedicineCare moved beyond the building.

    Control Health / Founder + CEO

    Built and exited a telemedicine platform in under two years, connecting patients and providers across messaging, video, billing, and compliance. Partnered with Zoom on establishing HIPAA-compliant communications protocol.

    • Telemedicine
    • Founder
    • Product analytics
  5. Connected health interfaceTouchscreens entered the sterile field.

    Inventor / U.S. Patent 9,550,620

    Invented devices and dispensers for sterile touchscreen covers, enabling tablets and smartphones to remain usable in clinical environments. Filed in 2012 and issued in 2017.

    • Healthcare
    • Touch interfaces
    • Patent
    U.S. Patent 9,550,620 (opens in a new tab)Filed January 31, 2012. Issued January 24, 2017.
  6. Mobile at scaleFrom message to measurable platform.

    Ping Mobile / SVP & CTO

    Built and scaled real-time mobile advertising infrastructure powering AT&T and Roku inventory. Attribution and lift systems delivered roughly 4x conversion gains and 50% re-engagement.

    • Mobile
    • RTB ad-tech
    • AT&T
    • Roku
    External record: Cisco author archive (opens in a new tab)Confirms SVP and CTO work spanning mobile platforms and connected health.
  7. Digital acquisitionPerformance became measurable.

    TFJ / Digital Marketing Director + Design Lead

    Led early digital acquisition and performance marketing across SEO and paid media, establishing foundational analytics and measurement strategy while digital commerce was still taking shape.

    • SEO
    • Paid media
    • Analytics
    • Experience design
05

Current questions

Questions worth keeping open.

The technology moves quickly. Better questions keep the work connected to reality.

  1. 01

    What should an intelligent system remember?

    Memory is product behavior, not infrastructure trivia.

  2. 02

    How do you know an agent did the right work?

    Quality should be inspectable before scale makes mistakes expensive.

  3. 03

    Where should autonomy stop?

    The human interface is also the accountability interface.

  4. 04

    Can a system improve without becoming opaque?

    Recursive learning matters only when the feedback loop remains legible.

06

Useful answers

In plain language.

What kind of AI product work does Isaac lead?

Isaac leads AI-native product strategy and execution across multi-agent orchestration, graph-based memory, knowledge architecture, evaluation systems, experimentation, and the human interfaces that make those systems useful.

What has Isaac shipped at scale?

Isaac has shipped products across AI, mobile advertising, e-commerce, fintech, telemedicine, and Apple platforms, including GoDaddy's first generative AI product, mobile infrastructure for AT&T and Roku inventory, and products recognized as Apple's Editor's Picks.

How does Isaac approach multi-agent product development?

Start with the outcome, design explicit roles and handoffs, make memory durable, instrument the system, contain the blast radius mechanically, and evaluate continuously.

What does Isaac focus on now?

Isaac focuses on building reliable AI product systems, helping teams adopt them without unnecessary jargon, and turning emerging capabilities into measurable advantages for customers and organizations.

View machine-readable profile

The full profile is available as JSON-LD and in a concise language-model summary.

Isaac Naor is an AI product leader in Los Angeles building multi-agent systems, durable memory, and evaluation frameworks into products that hold up after the demo.

Bio

The next useful thing

Hard problem?

Good. I'm most useful when the path is not obvious yet.

isaac@isaacnaor.comLinkedIn (opens in a new tab)