PortfolioNow: leading production for engine and tools teams on a AAA title

Kathryn Moore

Kathryn Moore

Discover. Scope. Deliver.

Product and delivery lead · Berlin

I turn messy problems into working solutions and products. I go into a team, find out what's really slowing it down, scope the fix with the people who'll use it, and see it through to release, building with AI when that's the quickest way. Fifteen years in product and delivery, often on internal tools and platforms other teams depend on, across AAA games, automotive and EV charging, a regulated health app, e-commerce and SaaS.

What I do

How I make a difference

Discovery and requirements

Turning a fuzzy idea into clear requirements and a plan a team can build from.

Where:EllimySugrMotesque

Roadmaps and prioritisation

Deciding what matters most, saying no clearly, and keeping the roadmap honest.

Where:UbisoftMercedes-Benz.ioDatarade

Delivery across teams and partners

Running work across teams, studios and vendors so leadership sees problems early.

Where:Ubisoft: six studios, three countriesMercedes me launchesImprove Digital

Operating models and ways of working

Setting up how teams plan, prioritise and hand over work, so it keeps running without me.

Where:UbisoftMotesqueDatarade

Regulated and complex products

Delivering where safety, compliance or many outside providers shape every decision.

Where:mySugr (medical)Elli (roaming and payments)Mercedes-Benz.io

AI in teams and builds

Bringing AI into teams with clear guidelines, and building small tools under a reviewed process.

Where:AI usage guidelinesAI goals in team OKRsMy own products
A recent example

Making the engine and tools teams dependable for a partner studio

The situation

A partner studio depends on the engine and tools teams. Requests arrived faster than they were answered, the tech leads were stuck on planning and tickets, and unplanned work kept breaking sprints.

What I changed

I rebuilt how requests come in and get prioritised, took planning and tickets off the tech leads, audited the teams' meetings, and introduced regular backlog reviews.

A first snapshot, three months in
  • The partner studio waits less. Requests rose by over half as the teams became more visible, yet more were answered within 30 days (up six percentage points), and the median wait more than halved.
  • The teams are more visible. Sprint planning updates for stakeholders, better triage, and a seat in the project's management meeting.
  • Unplanned work came under control. Unplanned work fell by a fifth, and every sprint stayed under the 25% limit.
  • More time goes into the roadmap. 59% more planned tasks a month, 30% more work overall.
  • The tech leads can focus on technical leadership. With planning and tickets off their plate, their completed planned work more than doubled, and about 15 hours a week went back into larger technical topics and expert support for the teams.
  • Nothing important sits forgotten. Bugs and crashes open for over 90 days fell by 62%.

Aug to Nov 2026 compared with Jan to Jul 2026, from the teams' own tracking data. The results are the teams'; my part was the changes above.

Judgement

Decisions I've made

The calls behind the results: what I chose, and why.

Ubisoft

Protecting planned work without shutting out the partner studio

The call
A 25% limit on unplanned work, kept for blockers, critical bugs and support for upcoming demos and gates. Everything else goes through triage.
Read moreShow less
The situation
Requests went straight into the sprint, so planned work was often started but not finished.
What happened
Unplanned work came under control, and the perception of the teams as a group that consistently delivers what was promised, when it was promised, vastly improved. The numbers are in the example above.
Ubisoft

Keeping bug priorities honest

The call
QA always own bug priority. Teams can challenge it and bring context to the QA team, but QA always make the final call.
Read moreShow less
The situation
Tech leads wanted to re-grade bugs QA had prioritised, which would have created a hidden second priority.
Why
QA see the whole project without bias and judge a bug by how much it hurts players or impacts the project in a holistic way. The team fixing it is naturally swayed by how much work the fix is, the stakeholder requesting it, their closeness to the issue etc.
What happened
A more collaborative relationship with QA, and much less time arguing about sprint priorities.
Ubisoft

Getting a shared view of pipeline quality

The call
Survey the tech teams and our main partners to get an evidence-based baseline before deciding what to fix.
Read moreShow less
The situation
How good our pipelines were depended on who you asked. Day-to-day feedback from the leads and our main partners told a different story from the project-level view.
What happened
A shared picture from both sides, with specific areas to act on. I shared it with leadership as a starting point.
mySugr

Knowing when not to ship

The call
Rework the flow through repeated UX and safety reviews until it was safe, and stop when the evidence showed it couldn't be.
Read moreShow less
The situation
I was product owner for a dose-calculation feature in a regulated diabetes app, where one confusing step could lead to a wrong dose.
What happened
It didn't launch. With patient safety at stake, all the work we had put in was no reason to release something that wasn't safe enough.
My own product

Numbers describe the work, never the person

The call
Numbers per team, goal and milestone only, never per person or ranked. Every spec asks whether anyone could feel watched.
Read moreShow less
The situation
A project management tool could easily show each person's speed, and some stakeholders would ask for it.
Why
It builds trust, and puts ownership on the team instead of blame on a person.
My own product

Stepping back from approving every screen

The call
Replace my sign-off with a clear standard and an independent design check. I'm only brought in when a main user journey is at risk.
Read moreShow less
The situation
I was approving every screen design myself before it was built, and I had become the bottleneck.
Why
My time is best spent on the calls only I can make.
What happened
Design work no longer waits for me, and quality is still checked on every screen.
How I work

Discovery to Delivery

This is how I build my own products, and what I enjoy setting up for teams, with or without AI.

  1. ResearchUnderstand the people, the data and the real problem before deciding anything.
  2. Requirements and scopeAgree what's in, what's out, and how we'll know it worked.
  3. SpecA written spec to a 12-part standard: walkthroughs, edge cases, failure handling, access rules.
  4. Separate reviewAn independent review checks each spec for gaps before anything is built.
  5. Build and testBuilt on a test branch, with unit and browser tests written alongside.
  6. Build reviewEvery problem is graded by severity, and serious ones are fixed before release. Risky changes are trialled first.
  7. Release and measureEach release is approved, then we check the result against what we wanted to accomplish.
Proof

Numbers from my own builds

The numbers come from a read-only audit of my most recent product's git history and documents, which I ran with an AI reviewer on 11 Nov 2026.

Delivery

40features shipped in 42 daysAudit 11 Nov 2026
1.9 daysmedian from spec to releaseAudit 11 Nov 2026
~90%of problems caught before releaseAudit 11 Nov 2026
5 of 5of the most serious problems fixed; 4 were caught before any code was writtenAudit 11 Nov 2026

Quality controls in my process

863unit tests, all passingAudit 11 Nov 2026
1,179browser testsAudit 11 Nov 2026
111 of 114tables with access rules (other 3 locked)Audit 11 Nov 2026
Experience

Where it comes from

  1. 2023 to now
    Ubisoft Berlin · Associate ProducerProduction for the engine and tools teams on a multisite AAA title, working across six studios in Germany, France and Canada; earlier, pillar producer on a multiplayer IP. Line manager across production and QA.
  2. 2021 to 2023
    Freelance product owner and delivery leadElli (Volkswagen Group): internal tools for EV charging networks. Datarade: a B2B data marketplace. Motesque: health tech, set up the working model from scratch.
  3. 2020 to 2021
    Mercedes-Benz.ioSenior product owner, Mercedes me companion app, 5M+ users.
  4. 2020
    NinoxSenior product owner on a low-code database product.
  5. 2018 to 2020
    mySugr (Roche)Senior product owner on a regulated diabetes app.
  6. 2012 to 2018
    BloomReach, Emakina, Improve Digital, Mando GroupBusiness analysis, QA and product roles in SaaS, e-commerce and advertising technology.

Certifications and education

  • PMI-ACPAgile Certified Practitioner, Project Management Institute
  • PSPO IProfessional Scrum Product Owner, Scrum.org
  • Leading SAFe 5.1Scaled Agile Framework
  • UX FoundationUser experience design
  • Risk management for medical devicesFor regulated health products
  • Leadership UnlockedReturn on Meaning, 2026
  • Ready to LeadLeadership programme, 2026
  • BSc Management and Information TechnologyLancaster University, UK

Learning now, with tools I built myself

  • GermanThrough my own German course, which I specified and built with AI around everyday life in Berlin: 20-minute lessons for my commute, with spaced-repetition review, listening and speaking practice, and vocabulary checked against the official Goethe A1 word list. The goal is the Goethe B1 certificate.
  • System Design for TPMsThrough a 10-module course I built with AI, combining the code of my own products with established outside sources such as the System Design Primer, AWS's Well-Architected Framework and Google's Site Reliability Engineering book, with exercises and a portfolio piece per module.
  • AWS Cloud Quest: Generative AI PractitionerIn progress. AWS's hands-on training on building with generative AI services, including how to keep them secure and responsible.
  • AWS Certified Cloud PractitionerPreparing now, as part of the system design course.
In their words

What people I've worked with say

“Our requirements were vague and dependencies unclear. Kat asked just the right questions to help us better understand our own requirements and dependencies.”

Dr. Brendan Koehler, EV charging, worked with Kat on the same team (2022)

“If there is a new feature request, before the task reaches the team, you could be 100% sure that this feature is consistent, [has] a design, description and acceptance criteria.”

Artem Skoretskiy, Team and Tech Lead, worked with Kat on the same team (2023)

“Kat was leading one of the most complex, time critical and important projects of the company … she ensured that the project would always stay on track, within scope and in the end be a success.”

Sarah Haaf, worked with Kat at mySugr (2020)

More recommendations on LinkedIn

Work with me

Two ways to work together

I find where your team loses time, agree the fix with you, and deliver it, as part of your team or on a project.

Employ me

Senior product, programme and delivery roles, especially for platform and internal tools teams, complex or regulated products, and work that spans several teams or companies. Berlin or remote in Germany.

Talk to me on LinkedIn

Hire me for a project

A defined piece of work, at a day or hourly rate: scoping a new tool, running or setting up a programme, or building it with AI.

€110 per hour
€850 per day

Your data and code stay in your tools. Nothing goes into AI models without your written agreement.

Get in touch on LinkedIn
Products

Products

Tools you can run on your own, without hiring me.

Coming soon

The Discovery to Delivery kit

My Discovery to Delivery way of building with AI, packaged for your team to run on its own: the 12-part spec template, the review checklists, the catalogue of ways AI-built software tends to fail, and the automated quality checks I use on my own builds.

Available as the full kit, or module by module if you only need one part, such as the spec template or the review checklists.