Data Scientist
Qogita
2 days ago
Full-time
On-site
London, United Kingdom
Data Scientist
Description
Wholesale is a $50 trillion market still working the way it did in 1950. 90% of transactions happen offline, over calls, catalogues and trade shows, and a product passes through four or five layers of distributors and wholesalers before it reaches a store, with every party deciding on a fraction of the picture. By the time it hits the shelf, it costs roughly 40% more than the maker charged. That inefficiency tax runs into the trillions, and everyone pays it every time they buy anything.
Qogita is building the operating system that replaces that chain: a single order book connecting the global market, the rails to move goods across borders, and a self-learning system that does every job once, against real-time data about the whole market. That puts data science at the core of the product. Forecasting reads what is actually selling in every market as it sells, a buyer's basket is assembled from every available seller and split across the optimal combination, and every trade writes a record that sharpens the next price, allocation and route. It is a hard problem, spanning physical goods, thirty national regimes in Europe alone and buyers who are increasingly agents, and it is already working: 95% of transactions involve no human touch, and the business is doubling every year.
The role
You're a data scientist with strong quantitative and ML chops who can move between classical modelling, experimentation, and modern applied ML (including LLMs where they're the right tool). You'll own end-to-end data science work across Qogita's wholesale marketplace: framing ambiguous commercial problems, building models that ship, and keeping them healthy in production. The Data Science team partners with Product, Engineering, Finance, and Commercial to build the intelligence layer behind pricing, matching, demand, and product discovery.
Wholesale is a $50 trillion market still working the way it did in 1950. 90% of transactions happen offline, over calls, catalogues and trade shows, and a product passes through four or five layers of distributors and wholesalers before it reaches a store, with every party deciding on a fraction of the picture. By the time it hits the shelf, it costs roughly 40% more than the maker charged. That inefficiency tax runs into the trillions, and everyone pays it every time they buy anything.
Qogita is building the operating system that replaces that chain: a single order book connecting the global market, the rails to move goods across borders, and a self-learning system that does every job once, against real-time data about the whole market. That puts data science at the core of the product. Forecasting reads what is actually selling in every market as it sells, a buyer's basket is assembled from every available seller and split across the optimal combination, and every trade writes a record that sharpens the next price, allocation and route. It is a hard problem, spanning physical goods, thirty national regimes in Europe alone and buyers who are increasingly agents, and it is already working: 95% of transactions involve no human touch, and the business is doubling every year.
The role
You're a data scientist with strong quantitative and ML chops who can move between classical modelling, experimentation, and modern applied ML (including LLMs where they're the right tool). You'll own end-to-end data science work across Qogita's wholesale marketplace: framing ambiguous commercial problems, building models that ship, and keeping them healthy in production. The Data Science team partners with Product, Engineering, Finance, and Commercial to build the intelligence layer behind pricing, matching, demand, and product discovery.
Requirements
- Build and deliver data science solutions across the stack: predictive models, ranking, demand forecasting, segmentation, pricing, experimentation, and LLM-powered features, depending on where the business need is greatest
- Take ownership of business-critical ML systems end-to-end: problem framing, model design, deployment, monitoring, and ongoing maintenance in production
- Translate ambiguous business problems into tractable ML or statistical problems with clear success criteria, working closely with Product and Commercial
- Apply quantitative methods (regression, causal inference, classical ML, and deep learning where useful) to pricing, demand, liquidity, supplier matching, catalogue enrichment, and related marketplace problems
- Design and analyse experiments and A/B tests, owning statistical validity and turning results into recommendations teams can act on
- Collaborate with Engineers to ship models via reproducible MLOps workflows: experiment tracking, model serving, alerting, and production monitoring
- Communicate model choices, limitations, and trade-offs clearly to both engineers and non-technical stakeholders