Istanbul-Based Operations Research Scientist

Arute Solutions; Fatih; On-site; TRY 900,000 - 1,300,000; Full time

Job summary

Arute Solutions in Istanbul is seeking an exceptional optimizer who loves turning hard optimization problems into production-ready models. You will spend most of your time building, refining, and maintaining mathematical models that steer decision-making at scale.

You'll design LP/IP solutions, work with CPLEX, Gurobi, and Hexaly, and implement production-grade code in C#/.NET and Python. Collaboration with AI/ML teams and translating real-world constraints into clean formulations are key.

Qualifications

• Strong background in operations research with LP/IP depth.
• Hands-on experience with CPLEX or Gurobi.
• Production-quality code around optimization models.
• Familiarity with AI/ML concepts and LLM agents.

Responsibilities

• Design, build, and maintain LP/IP models solving real operational problems.
• Work with CPLEX, Gurobi, Hexaly to formulate, tune, and troubleshoot models.
• Tackle routing, scheduling, and resource allocation problems.
• Develop heuristic approaches when exact methods are impractical.
• Write production-quality code around optimization models in C#/.NET.
• Prototype ideas in Python and review existing model code.
• Collaborate with AI/ML teams and LLM agents integration.
• Translate real-world constraints into clean mathematical formulations.

Skills

Operations research

Integer programming

Linear programming

Optimization

Python

C#

.NET

Gurobi

CPLEX

Math modeling

Communication

Tools

CPLEX

Gurobi

Hexaly

Job description

Arute Solutions in Istanbul is seeking an exceptional optimizer who loves turning hard optimization problems into production-ready models. You will spend most of your time building, refining, and maintaining mathematical models that steer decision-making at scale.

You'll design LP/IP solutions, work with CPLEX, Gurobi, and Hexaly, and implement production-grade code in C#/.NET and Python. Collaboration with AI/ML teams and translating real-world constraints into clean formulations are key.