# Andrew Durden

Director, Product Management, AI Experience Platform

Royal Caribbean Group, Miami · Since Jan 2017 · Director since Jul 2026

I lead Product Management for the AI Experience Platform and enterprise recommendations at Royal Caribbean Group alongside an extraordinary team of engineers, applied scientists, and designers. I own multi-year platform vision, capital cases, P&L, autonomous squad design, and Go-to-Market strategy across Web and React Native while staying close to live prototypes and evaluations. After hours, I teach PM cohorts at BrainStation, modernize forensic crash tools with my dad, and build smart-home hardware and games for family and friends.

- $300M/yr +13% cold-start ML booking conversion in A/B tests

- 110,000+/wk AI answer searches across 2 brands, up from 42K (95% helpful)

- $80M/yr omni-channel personalization across web, mobile & call center

Everything in this room was made for someone. Step inside the interactive 3D loft · Enter the room (6 MB).

Product Owner, E-Commerce Search & Booking (Jan 2017) → Senior Product Manager, Customer Intent (Oct 2021) → Senior Product Manager, E-Commerce Search & Personalization (Aug 2024) → Director, Product Management, AI Experience Platform (Jul 2026 to Present), all at Royal Caribbean Group, Miami; preceded by Zapporoo (Apr 2014 to Sep 2016).

## Work

Royal Caribbean Group

Jan 2017 to Present

### Director, Product Management, AI Experience Platform

Jul 2026 to Present

Own multi-year product vision, initiative P&L, capital allocation, and Go-to-Market execution alongside cross-functional engineering, applied science, and design squads building Royal Caribbean and Celebrity Cruises' shared consumer AI platform and enterprise recommendations across Web and React Native.

Multi-Year Vision, Capital Cases & P&L Ownership: Authored the 2026 follow-on and multi-year 2027 capital expenditure cases that won executive backing to unify isolated brand efforts into a shared platform that builds products; own initiative P&L, unit-economics modeling, executive reviews, and staged Go-to-Market rollout.

End-to-End Guest Journey & Compounding Data Strategy: Anchored roadmap bets in real guest friction from initial cruise discovery through pre-voyage planning and real-time onboard Net Promoter Score (NPS) recovery, treating structured guest preferences and context as a compounding multi-brand asset.

Build-to-Learn Discovery & Portfolio Prioritization: Shifted discovery from static spec handoffs to live interactive prototyping across Product, Design, and Engineering so squads test guest comprehension in days, double down on validated workflows, and cut low-impact scope before committing multi-quarter capacity.

Commercial Trade-Offs & Release Governance: Defined measurable launch gates grounded in real contact-center guest scenarios across answer quality, latency, and cost-per-conversation, aligning Applied Science, Legal, Privacy, and Brand teams around zero-regression release readiness and sunsetting underperforming features.

Autonomous Org Design, Blind-Review Hiring & Coaching: Designed the multi-squad 2027 operating model around end-to-end guest outcomes rather than component silos, instituted blind-review hiring across product roles, and coached product managers through promotion to Senior Product Manager.

### Senior Product Manager, E-Commerce Search & Personalization

Aug 2024 to Jul 2026

Direct-Answer AI Search Engine: Led cross-functional strategy and launch replacing legacy keyword search across both brand sites with a tool-calling AI answer bar (May 2026), growing usage from 42,000 at launch to 110,000+ weekly searches with a 95% guest helpfulness rating.

Support RAG Modernization & Platform Incubation: Rebuilt virtual-assistant knowledge retrieval (lifting self-service chat resolution without live-agent handoff from 49% to 56% and recall 68% to 78%), and built the early 2026 working prototype that secured multi-year AI platform funding.

Omni-Channel Personalization ($80M/yr) & Feature Platform: Delivered a shared omni-channel personalization platform generating $80M in incremental annual revenue across web, mobile, email, and call center, backed by a new enterprise Databricks real-time feature store.

### Senior Product Manager, Customer Intent

Oct 2021 to Aug 2024

0-to-1 Cold-Start ML Recommendations ($300M/yr): Founded the global marketing ML practice and replaced cold-start default Price: Low-to-High sorting with geographic and session-intent ranking, lifting booking conversion by 13% in controlled A/B tests and adding $300M in incremental annual booking revenue across web, mobile, and call center.

0-to-1 Saved Favorites, Watchlist & Multi-Model RAG: Launched Saved Favorites and Price Watch, turning anonymous cruise browsing into persistent cross-session watchlists, automated price-drop alerts, and first-party intent signals; migrated virtual-assistant flows to multi-model RAG (40% routine contact deflection) and introduced federated GraphQL.

### Product Owner, E-Commerce Search & Booking

Jan 2017 to Oct 2021

Award-Winning Search Grouping & Stateroom Booking: Shipped sailing-consolidation search grouping (+20% search-to-sailing funnel entry, MongoDB World 2019 Customer Experience Innovation Award), deployed the first production SageMaker recommender, and launched single-checkout connecting staterooms.

BrainStation

Concurrent Part-Time

Jan 2024 to Present

### Lead Product Management Instructor

Product Strategy & Discovery Cohorts: Teach 10-week certificate cohorts in Problem-Space discovery (5 Whys, Jobs-to-be-Done, Kano prioritization), Go-to-Market framing, live AI prototyping, and scope restraint.

Zapporoo

Apr 2014 to Sep 2016

### Design Manager & UX/UI Lead (Promoted from Mobile UX/UI Designer)

Consumer Mobile Product & UX Leadership: Advanced from hands-on mobile designer into design leadership guiding our Miami UX/UI team across 43 shipped consumer iOS and Android applications.

## Context Engineering & Builder Craft

Since building my first computers from scratch as a kid, I've never been able to use technology without wanting to open the case and understand why it behaves a certain way. Staying curious and getting hands-on is how I build to learn and partner with engineers far more talented than me: reading a slide deck never shows you where a system bends when jagged model behavior, latency, state, and human habits collide. When cloud primitives and coding agents let anyone generate code in hours, context engineering, systems fundamentals, and architectural matching matter far more than commit counts:

Dual-Plane Context Engineering (Runtime + Harness) Treating the context window like scarce working memory across both customer-facing agents (Write to scratchpads for +10 percentage-point lift in multi-turn task completion, Select from structured context graphs and non-overlapping MCP tools, Compress via multi-thread compaction and tool-result clearing, and Isolate across sub-agents) and team coding harnesses (concise CLAUDE.md / AGENTS.md conventions, on-demand SKILL.md workflows, and deterministic hooks where probabilistic rules are unsafe).

Caching, Key-Value State & Security Boundaries Architecting multi-tier state across distributed key-value stores, Databricks feature caches, and LLM prompt caches (preserving prefix stability across dynamic tool loading for 12% to 22% lower LLM inference cost per guest conversation at flat latency while enforcing strict KV TTL invalidation, session isolation, and PII redaction so live pricing or guest state never goes stale or bleeds across sessions).

Problem-to-Architecture Matching, Evals & Restraint Matching each human problem to the right computing layer rather than forcing LLMs everywhere, gating probabilistic agents with golden eval datasets and calibrated LLM-as-judge CI while keeping CrashCalc's highway physics 100% deterministic and offline-first on SQLite, keeping Messina OS local on a home Raspberry Pi LAN, and sunsetting ProductPeople the week native platform memory made a standalone wrapper redundant.

## Projects & Personal Builds

Building CrashCalc: Rebuilding my father's commercial forensic crash-reconstruction software in Kotlin Multiplatform (one audited physics engine across native iOS and Android): 6 pinned roadside tools, live phone slope sensor, printed textbook citations (Fricke 2010; Daily & IPTM 2018), and offline resilience so a paywall never appears at a crash scene.

Sunset Mar 2026 ProductPeople · 2025 to 2026: Customer-interview memory workspace built for PMs to pressure-test decisions against real transcripts. Piloted with cross-industry peers in late 2025 and cleanly retired in March 2026 once frontier tools shipped native project memory, carrying forward its spec-driven eval gates.

Building Messina OS: Built for our household and our two Goldendoodles, Obi and Chewy. Unifies a fragmented smart home across idle 65-inch Samsung TVs via a Raspberry Pi hub and native Swift iPhone app: room-by-room Sonos control, live security cameras on any screen, zero banner ads, Playwright visual diffs against Figma with instant undo, and remote Dog Mode (dimming lamps and playing warm jazz even when we are away).

Shipped The Wedding Camera & Hidden 16-Bit Platformer: Zero-install browser camera built for guests at [Casa Tua](https://www.casatualife.com/Miami.html) in Miami (6 camera bodies, 14 custom candlelight film looks), hiding a secret 16-bit Miami platformer and scavenger hunt underneath the clean shutter so relatives saw one button while curious best friends unlocked the game.

Shipped Bilingual Legal Web Platform & Custom MCP Server (2025): Shipped a bilingual English/Spanish legal platform backed by a headless CMS, 8 Playwright end-to-end test suites, and a custom MCP server for AI client intake.

## Teaching

Internal AI Enablement · Royal Caribbean Group: Coach PMs, designers, and engineers at my desk, in workshops, and through shared plugins to prototype working software early and test with real customers before committing engineering roadmaps.

## Skills & Education

**Product Strategy & Commercial Leadership:**

- Multi-Year Vision & Roadmapping,

- P&L & Unit-Economics Ownership,

- Capital Expenditure Business Cases,

- Executive Storytelling,

- Go-to-Market (GTM) Sequencing,

- Funnel & Pricing Optimization

**Discovery, Org Design & Execution:**

- Jobs-to-be-Done (JTBD) & Kano Prioritization,

- Build-to-Learn Prototyping,

- Controlled A/B Experimentation,

- Autonomous Squad Topology,

- Blind-Review Hiring,

- PM Coaching & Career Ladders

**AI, Personalization & Technical Fluency:**

- Consumer AI Platforms,

- Cold-Start ML Ranking & Search,

- Feature Engineering,

- Evaluation-Gated Release Governance,

- Context Architecture,

- Web & React Native,

- SQL,

- Python,

- TypeScript,

- Figma & Adobe CC

**Education:**

- Florida International University, Miami: undergraduate studies.

## At Home

At home I build tools for our household, our wedding guests, my closest friends, and my father's forensic practice.

I've been a tinkerer since early childhood, gaming, building PCs on the carpet, and taking things apart just to see why they worked. Putting on the blue polo at Best Buy during FIU felt like a dream job, and that curiosity never left.

Weekends revolve around our backyard pizza oven and 48-hour cold-fermented NY sourdough (64% hydration).

House pour at the bar: an equal-parts Paper Plane with Bourbon, Aperol, Amaro Nonino, and fresh lemon.

Two Goldendoodles run the house, Obi and his son Chewy.

I play tenor saxophone, starting in high school marching band and playing packed Miami ska shows with The Skandals.

I still hide Easter eggs in almost everything I build for friends to find.

## Contact & Resume Downloads

Building consumer AI, evaluating leaders, or want to collaborate? Send a note below or connect on LinkedIn.

[LinkedIn](https://www.linkedin.com/in/adurd001)

Resume PDFs

- [Product Leadership](../resumes/Andrew-Durden_2026-Resume-Product-Leadership.pdf)

- [Builder](../resumes/Andrew-Durden_2026-Resume-Builder.pdf)

- [Design](../resumes/Andrew-Durden_2026-Resume-Design.pdf)

- [AI Product Manager](../resumes/Andrew-Durden_2026-Resume-AI-Product-Manager.pdf)

- Work

- Context & Craft

- Projects

- Teaching

- Contact

- [Enter 3D Loft →](../index.html?enter)

Miami · The 3D loft is an illustrated space, not a photograph of our home · Nobody is named or pictured without permission.

[SKILL.md](../SKILL.md) · [llms.txt](../llms.txt)
