Senior Data Scientist - (Content, Consumer)

Delivery Hero
Berlin
Workplace: HybridFull timeFunction: Data Science & Machine LearningSkills: ["Ownership","Mentoring","Product thinking","Problem framing","Pragmatism"]

Own end-to-end LLM systems powering ratings and reviews (“social proof”) across many markets and languages. Define the technical direction for model/provider selection, prompt strategy, and offline/LLM-judge evaluation infrastructure, including drift and quality monitoring in production. Drive the roadmap by discovering high-impact gaps, quantifying business value, and taking prototypes into production with backend and data engineering.

Loading

Loading job details...

Preparing the role view and application actions.

FursaFursa
Delivery Hero
Delivery Hero
1 day ago

Senior Data Scientist - (Content, Consumer)

✓ Verified Job

Canonical indexed version, validated from employer's careers page.

Source: Company careers pageValidated by: Fursa AI
Last checked: 2 hours agoStatus: Live
Reposted: similar role first listed 2 months ago

Job Summary

Own end-to-end LLM systems powering ratings and reviews (“social proof”) across many markets and languages. Define the technical direction for model/provider selection, prompt strategy, and offline/LLM-judge evaluation infrastructure, including drift and quality monitoring in production. Drive the roadmap by discovering high-impact gaps, quantifying business value, and taking prototypes into production with backend and data engineering.
Location: Berlin
Workplace: Hybrid
Employment Type: Full time · Permanent
Job Function: Data Science & Machine Learning
Seniority: Mid level

Key Responsibilities

  • •Own the LLM systems behind social proof end-to-end, ensuring quality, reliability, and coverage across languages, platforms, and use cases, including production drift and data issues.
  • •Set the technical standard for how the team builds with LLMs: model/provider selection using empirical cost/latency/quality evidence and systematic prompt strategy.
  • •Drive the roadmap by discovering problems, quantifying business value, and turning them into scoped initiatives—treating inference cost as a product decision.
  • •Define and operationalise evaluation metrics and infrastructure (offline eval suites, LLM-as-judge frameworks, annotation processes) so changes can be assessed quickly.
  • •Take prototypes to production by shaping architecture and data flows with backend and data engineering, and building feedback loops to improve without constant manual intervention.

Pay and Benefits

Equity and Bonus:Equity
Perks:Educational BudgetLanguage CoursesParental SupportHealth InsuranceGym MembershipLife InsuranceCorporate PensionMeal VouchersCommuter Benefits

Key Requirements

  • •Built LLM-based systems that reached production, preferably on noisy, multilingual, user-generated text, with the ability to explain architecture and failure modes.
  • •Designed offline evaluation suites for generative systems and used LLM-as-judge patterns with awareness of bias, measuring faithfulness, hallucination, and quality at scale.
  • •Instrumented LLM pipelines with observability (e.g., tracing chained calls, tracking token usage/cost, capturing intermediate outputs) for multi-step debugging.
  • •Applied strong statistics foundations and causal inference, including challenging results that look too good.
  • •Proficient with production-grade Python and analytical SQL, with familiarity across ML lifecycle practices, monitoring, orchestration, and engineering standards like version control and CI/CD.
Skills:OwnershipMentoringProduct thinkingProblem framingPragmatism
Languages:English
Tech Stack:PythonSQLLLMNLPCI/CD

Company Brief

Delivery Hero
Global online food-ordering and local delivery platform operating multiple regional brands (e.g., Foodpanda, Glovo, Talabat) across 50–70+ countries, supplying food and quick-commerce deliveries via marketplace and logistics services.
Industry: Online Marketplaces
Company Size: Enterprise (1,001+ employees)
Revenue: USD 1B+
Growth: Public Company
Valuation: Public Company (Market Cap in USD)
Funding: IPO / Publicly Listed
Headquarters: Berlin, Germany
Founded: 2011
Glassdoor
Glassdoor: 3.4
WebsiteLinkedInGlassdoor