Qogita logo

Data Scientist (LLM)

Qogita
1 day ago
Full-time
On-site
London, United Kingdom
Data Scientist
Description

You're a data scientist with broad analytical and ML experience as well as production LLM expertise. You'll own the full spectrum of data science work at Qogita β€” from classical modelling and forecasting through to LLM-powered features β€” and act as the team's go-to on language model architecture, evaluation, and deployment. You'll take end-to-end ownership of complex ML systems and pipelines that are business-critical: designing them, shipping them, and keeping them healthy in production. The Data Science team works cross-functionally with Product, Engineering, and Commercial teams to build the intelligence layer that drives Qogita's marketplace.

Requirements

  • Build and deliver data science solutions across the stack β€” predictive models, ranking systems, demand forecasting, and LLM-powered features β€” depending on where the business need is greatest
  • Take ownership of business-critical ML systems end-to-end: from problem framing and model design through to deployment, monitoring, and ongoing maintenance in production environments
  • Act as the team's domain expert on LLMs: advise on model selection, architecture decisions, prompt engineering, fine-tuning, and evaluation
  • Design and implement RAG architectures and evaluation frameworks where language models are the right tool for the problem
  • Apply classical ML and statistical modelling to structured business problems β€” pricing signals, supplier matching, catalogue enrichment β€” with rigorous attention to measurement and validation
  • Translate ambiguous business problems into tractable ML problems with clear success criteria, working closely with Product and Commercial stakeholders
  • Collaborate with Engineers to ship models via reproducible MLOps workflows β€” experiment tracking, model serving, alerting, and production monitoring β€” with a high bar for reliability and observability
  • Communicate model choices, limitations, and trade-offs clearly to non-technical stakeholders including Product and commercial leadership