001✓ copiedabout
About
Clemence W. Chee is an independent data and AI executive in Berlin. He shipped the first AI agents at Babbel and built data functions from zero at HelloFresh, and now works as an interim CTO for companies whose product is data or AI, and as a Chief Data & AI Officer for companies that already have scale.
I started in geodesy at TUM, reconstructing a city in four dimensions from satellite radar data. That trained a specific reflex: before you model anything, earn the right to trust your measurements. Everything since has been the same move at organisational scale.
The systems I built have mostly been replaced by now. The people I hired are running data functions across Europe. That is the part I am proud of.
I coach, and I build. Those are the two things I am actually good at.
I do not run general platform engineering, infrastructure, security or SRE.
002✓ copiedtimeline
Timeline
End date pending. Left Babbel by July 2026.
Vice President, Data & Analytics
Babbel GmbH · Berlin, Germany
Owned data and analytics at Babbel: the platform, the analytics function, a data product management team that did not exist before 2023, and the company's first AI agents in production.
- I led the data transformation at Babbel. The first piece of it was a set of training modules every new data hire worked through in their first weeks.40% time to onboard a new data hire
- Babbel had no data product management function before 2023. I wrote the role, built the hiring loop, and staffed the team myself.7 people data product managers
- Revenue reporting at Babbel ran on several competing data models. I standardised them into one set, and financial forecasting was rebuilt on top of it.95% availability of the standardised revenue data models
- Storage was the largest line in the data platform budget. I cut it and put the money back into platform work.60% data platform storage cost
- Answering a question at Babbel took two to four weeks. I rebuilt the platform underneath it so the same questions resolve while you are still in the meeting that raised them.10 minutes time from question to delivered insight
- Babbel had no AI agents in production. I shipped the first one, an internal application, and then the customer-facing agents that followed it.0 systems AI agents in production before this work
- I kick-started the AI transformation by putting AI agents in the hands of engineers first, so the organisation learned on internal tooling before anything reached a customer.
- I ran AI compliance with Legal as a partner rather than as an approval gate, covering internal policy, the EU AI Act and GDPR.
Global Senior Director, Data
HelloFresh SE · Berlin, Germany
Ran data for HelloFresh globally: the strategy, a new data management department, the Data Academy, and the budget behind all three.
- I wrote the global data strategy for HelloFresh and ran it across every market the company operated in.50% operational efficiency under the global data strategy
- HelloFresh had no data management department. I founded one and gave it governance and process automation as its mandate.more than 150000000 USD data risk exposure brought under management
- Data quality, data education and data management had no dedicated teams when I took the role. I hired into all three from a standing start.16 people people in the data management department
- I founded the HelloFresh Data Academy so people outside the data team could learn to read and build with company data.more than 100 people employees certified through the Data Academy
- I moved machine learning into business processes at HelloFresh and set up the innovation labs that carried the research behind them.
- I owned the data budget and allocated it against value-based OKRs, so every line in it had an outcome attached.
Senior Director, Operations Business Intelligence
HelloFresh SE · Berlin, Germany
Operations business intelligence for HelloFresh worldwide. The first data product, the shared KPI set for operations and finance, and a rebuilt data architecture.
- I built the first data product HelloFresh shipped. It automated work the operations teams had been doing by hand.more than 2000 hours per year manual work removed by the first data product
- I moved machine learning out of analysis and into the operations decisions that run the network day to day.more than 1% profit contribution margin, group level
- Operations and finance measured the business with different definitions. I wrote one financial and operational KPI set and made it the basis for the weekly decisions.
- I rebuilt the operations data architecture. Analysts stopped waiting on the old one to answer a question.1 week time from question to delivered insight
Director, Operations Business Intelligence
HelloFresh SE · Berlin, Germany
Founded the internal business intelligence centre of excellence at HelloFresh and led the architecture and supply chain modelling work underneath it.
- HelloFresh had no central home for business intelligence. I founded the internal centre of excellence and grew it into one.20 people people in the business intelligence centre of excellence
- I led the data architecture rebuild that every downstream reporting pipeline was then moved onto.50% data pipeline efficiency
- I built the supply chain models that procurement and planning used to manage cost of goods sold.8% product margin under the supply chain models
Head of Business Intelligence & Business Development
HelloFresh Australia · Sydney, Australia
First analytics hire for HelloFresh Australia. Built the team, the finance and procurement data systems, and the reporting the market ran on.
- I designed the finance and procurement data systems for the Australian business during its first period of scale.more than 1000000 USD cost saved by the finance and procurement data systems
- Australia had no analytics capability when I arrived. I hired the first business intelligence team there and ran it.4 people people in the Australian business intelligence team
- Procurement data was unreliable enough to distort the forecast. I rebuilt the accuracy checks behind it.0.02% procurement data error rate
- I put machine learning into the operational decisions the Australian business ran on, close enough to the P&L that the effect showed up in contribution margin.
Global Venture Development Manager, Head of Operations
Rocket Internet SE, ShopWings · Manila, Philippines
Ran operations for ShopWings in the Philippines through launch, covering the market entry analysis and the retail partnerships behind supply.
- I built the market entry analysis the Philippines launch was decided on.
- I set up the operating framework the retail partnerships ran on.
Global Venture Development Manager, Operations
Rocket Internet SE, ShopWings · Munich and Berlin, Germany
Venture development for ShopWings out of Munich and Berlin, building the geospatial and capacity models the delivery operation was planned with.
- I built the routing and delivery data products in ArcGIS, Python and R.14% margin on deliveries covered by the routing models
- I built the capacity planning models the supplier network was scheduled with.more than 100 suppliers suppliers covered by the capacity planning models
Education: Bachelor in Geodesy and Geoinformation, Technical University of Munich, 2009 to 2013, plus one completed master semester.