Skip to content

AI Product Manager

Operations-heavy AI products.

Enterprise hotel operations, hands-on AI product builds, evaluation loops, human approval, and rollout work for systems that have to hold up under real operating pressure.

Reztrix shows the product build. MGM shows the operating scale. The supporting work shows control, review queues, and audit trails for AI inside messy workflows.

Reztrix proofContact
18
years in hotel operations

MGM Resorts hotel operations

42% to 84%
Reztrix eval gain

50 synthetic Reztrix scenarios

15% to 32%
mobile adoption lift

9,000+ room rollout support

Evidence

Reztrix, MGM operations, and AI controls.

The strongest evidence is Reztrix and MGM. Support projects add governance, regulated review, and workflow-control context without making every project carry the same weight.

Primary evidence

Reztrix, MGM, adoption

  • Reztrix AI hospitality operations prototype
  • MGM enterprise hotel operations leadership
  • MGM AI guest-support rollout
  • Mobile check-in adoption rollout
  • OPERA property management system (PMS), user acceptance testing (UAT), requirements, and rollout work

Control evidence

Review, refusal, audit

  • Spatial Limen refusal-first governance and provenance
  • ACATS review queue and audit workflow
  • Contract Change scope-control prototype
  • Human-in-the-loop automation kits
  • AI, product, cloud, and Scrum credentials

Role-dependent support

Domain-specific product work

  • TeeTimeBuddy for consumer workflow/product execution
  • Studio Ops for owner-led operations consoles
  • BurgReport for public-data product work
  • 702 Turf for local conversion systems
  • MBA and hospitality degree as background context

Background only

Accurate, lower signal

  • Generic cloud certification collecting
  • Low-level DevOps or infra positioning
  • Every project screenshot at equal weight
  • Broad consulting offer copy
  • Unqualified product ownership claims for enterprise systems

Product proof

Reztrix, MGM rollout work, and control loops.

Flagship build

Reztrix - RAG, evaluation, and manager review.

Reztrix is an independent AI hospitality operations prototype for policy-grounded recommendations, structured outputs, manager review, and eval-based iteration.

  • Retrieval-augmented generation (RAG) grounded in approved operating documents, with 42 standard operating procedures (SOPs) and 225 embedded chunks.
  • Model-graded evaluation across 50 synthetic frontline scenarios, with answers scored against an explicit rubric.
  • Safety-oriented pass rate improved from 42% to 84% over 10 iteration cycles.
  • Manager approval keeps recommendations in decision-support mode, not autonomous action.
Evidence of requirements, product logic, evaluation design, prioritization, and reviewable workflows for hotel operations.
View proof

Enterprise adoption

MGM - enterprise operations and adoption.

Eighteen years in MGM hotel operations, including Executive Director of Hotel Operations at MGM Grand, grounding product decisions in frontline reality and executive constraints.

  • Operational lead for the AI guest-support assistant rollout across guest services, phone switchboards (PBX), and concierge teams.
  • 20% inbound call-volume reduction from the AI guest-support rollout.
  • Business-side mobile check-in rollout support across 9,000+ hotel rooms.
  • Increased mobile check-in adoption from 15% to 32%.
Evidence of user acceptance testing, vendor and stakeholder coordination, frontline enablement, and measurable adoption.

Controls and governance

AI controls - refusal, provenance, and audit trails.

Spatial Limen, ACATS (the regulated US process for transferring a customer account from one brokerage to another), Contract Change, and automation work show product judgment around evidence, refusal, source provenance, approval states, and auditability.

  • Spatial Limen uses refusal-first review gates across 27K+ research nodes, where weak matches are rejected by default and only reviewed work is shown.
  • ACATS Workbench supports regulated review queues, feed coverage, pinned decision facts, and audit logs.
  • Contract Change maps intake, scope, approval, and traceability states.
  • Automation kits use classify, draft, approve, act, and log patterns.
Shows the judgment AI PMs need: what a model can draft, what rules can route, and what a person must approve.
View proof

Work surfaces

Product surfaces worth opening.

Reztrix first, followed by governance and regulated-workflow examples with explicit prototype and synthetic-data boundaries.

Competencies

Requirements, evals, approval, rollout.

Requirements and workflow translation

MGM operations, OPERA/PMS work, Reztrix workflow design

Turns operating friction into product requirements, acceptance criteria, and product flows.

AI evaluation discipline

Reztrix 50 synthetic scenarios, model-graded eval loop

Defines tests before trusting model output and improves against explicit synthetic scenarios.

Human approval and control design

Reztrix, automation kits, Contract Change, ACATS

Separates model drafts, rule-based routing, and human decisions.

Enterprise rollout and user acceptance testing

MGM AI rollout, mobile check-in, OPERA planning/testing

Understands launch risk, training, vendor coordination, user acceptance, and frontline adoption.

Operational prioritization

Reztrix guest-impact scoring, MGM service operations

Designs ranking and escalation logic for time-sensitive service work.

Role targets

Operations-heavy AI PM roles.

  • AI Product Manager for operations-heavy internal tools or service workflows
  • AI Product Owner for workflow automation, contact center AI, hospitality tech, or internal operations tools
  • Technical Product Manager for AI workflows where adoption and governance matter more than model research
  • Forward-Deployed AI Product Manager for discovery, prototype translation, and customer workflow design
  • Product Operations or AI Enablement Manager owning process, tooling, evals, and rollout

Boundaries

Prototype, rollout, and ownership limits.

  • Reztrix is an independent prototype, not production software, and its evals use synthetic scenarios.
  • MGM AI assistant work is operational lead and rollout support, not end-to-end product ownership.
  • Mobile check-in is business-side rollout support, UAT coordination, vendor coordination, and frontline enablement.
  • Spatial Limen support proof uses 27K+ research nodes plus refusal and provenance framing.
  • Credentials support the role signal but do not replace the product and rollout evidence.

Contact

Talk about an AI PM role.

Send a note, use the email path, or inspect the structured resume data. The same proof is available through the work links above.

Resume JSON