Staff Data Scientist I - Matching careem

As a Staff Data Scientist I Matching, you will own the end-to-end matching and dispatch architecture for Careem s Food and Groceries verticals. You will design and implement real-time assignment systems that optimize cost, SLA compliance, and marketplace efficiency at scale.

This role sits at the core of the marketplace engine solving complex, multi-objective optimization problems under uncertainty. You will build systems that dynamically match orders to captains, optimize batching and pooling decisions, and continuously re-optimize as new information arrives. You will operate as a senior technical IC, leading architecture decisions, mentoring other data scientists, and setting optimization standards across the domain.

What You'll Do

1. Matching & Dispatch Architecture Ownership
   
   • Design and own the end-to-end matching and dispatch systems for Food & Groceries.
   • Architect scalable assignment frameworks incorporating:
   • Static and dynamic assignment
   • Online matching under uncertainty
   • Batch and pooling optimization
   • Continuous re-optimization with dynamic events
   • Establish robust optimization standards and best practices across the organization.
2. Multi-Objective Optimization & Algorithm Development
   
   • Formulate and implement multi-objective optimization models balancing:
   • Cost minimization
   • SLA compliance
   • Marketplace efficiency
   • Develop and deploy:
   • Min-cost flow and assignment models
   • Graph-based optimization frameworks
   • Custom solvers and OR-Tools based solutions
   • Incorporate stochastic elements and uncertainty-aware decision-making into assignment policies.
   • Continuously refine matching quality under high-scale marketplace conditions.
3. Real-Time Systems & Production Delivery
   
   • Build production-grade real-time matching systems using Python, Spark, and Trino.
   • Design scalable pipelines capable of handling large-scale, high-throughput marketplace events.
   • Collaborate with Engineering and Platform teams to ensure reliable model serving and system performance.
   • Lead performance monitoring, diagnostics, and system optimization.
4. Experimentation & Impact Measurement
   
   • Define evaluation metrics aligned with marketplace and operational objectives.
   • Design and run controlled experiments to quantify matching improvements.
   • Analyze system trade-offs and communicate clear, data-driven recommendations to Product and Operations stakeholders.
   • Ensure decisions translate into measurable improvements in efficiency and SLA performance.