I design the software people rely on at work, and the AI that goes into it.
- Now
- Product Lead at PackBack Network, a Dutch smart-reuse system for food and drink packaging.
- Before
- Five years at Schneider Electric, from product designer to digital product lead.
- On the side
- Building Langwish, my own iPhone app, and studying Computer Science for AI at Harvard, online.
PackBack Network
Product Lead · Nov 2025 to now
PackBack replaces disposable cups and food packaging with reusables that people borrow from a SmartDispenser and drop into a SmartBin. I lead product across the machines' screens, the PackPulse dashboard for operators, the field crew's iOS app and the design system that ties them together.
PackPulse Insights
Jan to Jun 2026A redesign of PackBack's operator dashboard around a short weekly list of AI findings. Each one says what changed at a location, why we think so, and what to do about it.
Result: Locations that acted on insights raised their return rate by 6.8 points
SmartBin and dispenser screens
Feb to May 2026The screens, lights and sounds on PackBack's return machines and the new hands-free dispenser. Designed for someone in a hurry, holding a cup, who has never used one before.
Result: Median return time from 9.8 s to 4.6 s, failed returns down 58%
PackBack Ops for iOS
Mar to Aug 2026A native iOS app for the crew that keeps the reuse system running: restocking dispensers, emptying SmartBins, fixing machines and installing new locations. It works in basements with no signal.
Result: Stops per route up from 9 to 13, machine downtime down 44%
Loop design system
Nov 2025 to nowOne design system for three very different surfaces: the PackPulse web app, the crew's iOS app and the touchscreens on our machines. Built with Arabic and right-to-left layouts from the start, for the Gulf launch.
Result: 84% of UI now built from Loop, Arabic screens shipped in 5 weeks
PackPerks
May to Sep 2026A rewards programme for returning reusable cups, designed around one hard rule: no app and no sign-up. Your contactless card is the account, and progress lives in your phone's wallet.
Result: Return rate up 7 points at pilot locations, 0 apps downloaded
Impact Calculator and rollout planner
Dec 2025 to Feb 2026An honest impact estimate on the website for prospects, and an internal planner where AI suggests where machines should go and turns it into a proposal. Sales went from weeks of spreadsheets to a two-day turnaround.
Result: Proposal turnaround from 15 days to 2, demo requests up 2.3×
Schneider Electric
Product Designer, then Digital Product Lead · Jul 2018 to Jul 2025
Five years on software for the people who keep buildings, factories and fleets running: facility managers, installers, panel builders and field technicians. I joined as a product designer and spent the last four years as Digital Product Lead, owning strategy, AI integration and process automation for a group of B2B tools.
Atlas Copilot
2023 to 2024An AI assistant inside a building-management app. Facility managers ask questions about their buildings' energy data and get answers with sources, a confidence level and drafted actions that a person approves.
Result: Time to diagnose an alarm down from 26 to 16 minutes
Alarm Inbox
2024 to 2025Turning hundreds of building alarms a day into a short list of incidents. Related alarms are grouped, ranked with a reason, and muted only with an expiry date and a note.
Result: 340 alarms a day became 41 incidents, with no critical alarm missed
Commissioning Assistant
2024 to 2025AI that reads a building's electrical drawings and network scans and proposes the whole monitoring setup. Installers review only the points the model isn't sure about.
Result: Site setup from 3 days to 4 hours across the first 120 sites
Quote-to-Order
2022 to 2023Replacing a legacy configurator, Excel sheets and long email threads with one guided flow for partners who order electrical switchboard equipment. Problems get caught while configuring, and standard quotes approve themselves.
Result: Time to a valid quote from 2 days to 25 minutes
EV depot charging planner
2021 to 2022A web planner for overnight charging of electric buses and vans at depots. Every vehicle is ready by its departure time, without going over the site's grid limit.
Result: Vehicles not ready at departure from 6.2% to 0.8%
FieldPro for iOS
2019 to 2021An offline-first iOS app for field technicians who maintain switchgear and UPS systems at customer sites, often in basements with no signal and gloves on.
Result: First-time fix rate up 18%, report time down 45%
Concepts and own work
Concepts and my own app · 2024 to now
Self-initiated concept projects in the real visual identity of companies whose products I use: bol, Vattenfall, Zonneplan and Back Market. They are unofficial and not affiliated with those companies. I do them to practise on problems outside my day job. Plus Langwish, my own iPhone app.
Returns that sell again
bol · ConceptA concept for the returns section of bol's Partnerplatform. AI reads the buyer's reason and photos, suggests whether an item goes back on sale, gets repaired, is sold as Tweedekans or recycled, and the seller approves with one click.
Result: In tests, sellers handled ten returns in 4 minutes instead of 11
One screen for 'why is my bill higher?'
Vattenfall · ConceptA concept for Vattenfall's customer service agents. One workspace that explains a customer's annual settlement from their meter data and drafts a reply the agent edits and sends, with every fact linked to its source.
Result: In tests, answering a billing email took 6:10 minutes instead of 9:40
Quoting a roof before the ladder comes out
Zonneplan · ConceptA concept for Zonneplan's solar advisors and planners. AI proposes a panel layout from aerial data with shading shown up front, the advisor adjusts it and sends a quote the same day, and one roof visit confirms everything before installation.
Result: In tests, a first quote took 6 minutes instead of 35
Making 'Excellent' mean the same everywhere
Back Market · ConceptA concept for a grading assistant Back Market could give its refurbisher partners. A photo booth and AI mark scratches, diagnostics run automatically, and the technician confirms the grade, so 'Good' or 'Excellent' means the same at every refurbisher.
Result: In a blind test, technicians agreed on the grade 88% of the time, up from 58%
How I work
- 01
Name the outcome first
Before any screen I want two things: the number we are trying to move, and the person whose day gets better. “Cut the time it takes to diagnose an alarm” leads somewhere better than “add a chatbot”.
- 02
Design the bad day
Declined cards, full bins, no signal in the basement, a model that isn’t sure. I map those states before the happy path, because that is where people decide whether to trust a product.
- 03
Keep people in charge of the AI
Show the source, show how sure it is, and turn actions into drafts someone approves. If a person is accountable for the result, the interface has to let them check it.
- 04
Prototype in the real material
Real data, the real device, the real place. I prototype in code when that is the fastest way to learn, and I test machines where they are installed.
Experience
Education
About
I trained as an industrial design engineer at The Hague University. That meant learning about tolerances, materials and factory constraints long before I opened Figma, and it left me with a habit I still have: I want to know how a thing is built before I decide how it should look.
At Schneider Electric I started on software for facility teams and field technicians, then became Digital Product Lead for a group of B2B tools. That is where I learned to work on AI features with data scientists, to automate processes that ran on email and Excel, and to defend a roadmap in front of people who sign the budget.
Since November 2025 I lead product at PackBack Network. The work is unusually physical for software: a touchscreen on a return machine, a dispenser that has to hand you a cup in a few seconds, an app for the crew who empties the bins. I like that a lot.
- Languages
- English, Dutch, Arabic, Russian
- Good at
- Product design, AI integration, Design systems, UX and UI research, Software engineering, Strategy, Team leadership, Graphic design
- Tools I use most
- Figma, FigJam, ProtoPie, Framer, React and TypeScript, Cursor, Claude and OpenAI APIs, Storybook, Maze, Dovetail, Amplitude, Linear, Jira



