Math-Driven, Business-Focused

Turning your optimization challenges into decision support software

Guillaume Marques, PhD — Freelance Operations Research Consultant. 10 years of experience. I model, I code, I industrialize. From requirements to production deployment.

What I do

I solve linear, non-linear, discrete, and continuous optimization problems. My primary focus is on Energy, Logistics, and Mobility.

Custom Mathematical Modeling

Don't settle for “off-the-shelf” solutions that don't meet your needs. I formalize your company's specific constraints into mathematical models. I leverage large-scale decomposition techniques, column generation, and matheuristics to solve your most complex instances.

Performance-First Development

I work with both open-source (HiGHS, Ipopt, SCIP) and commercial engines (Gurobi, CPLEX, Knitro). My code is not just a prototype; it is production-grade software built to tackle your most critical operational challenges.

Production-Ready Engineering

Mathematical models are useless if they don't integrate. I provide full-stack industrialization: CI/CD, robust APIs, and cloud architecture to ensure your optimization engine is a reliable asset, not just another piece of shelfware.

How I work

I offer both time-and-materials (acting as an embedded expert for a specific project) and fixed-price project delivery.

01

Scope

Understanding your business problem, identifying constraints, and defining optimization levers.

02

Model

Translating the problem into a mathematical model. Everything is fully documented.

03

Develop

Coding the resolution algorithm tailored to your time and quality requirements (exact methods or heuristics).

04

Integrate

Designing the architecture to integrate the software into your IT system (API, database, deployment).

05

Validate & Handover

Defining the validation strategy, writing technical docs, and training your teams.

Case Studies

GRDF — French national gas network

Developing mission-critical optimization and simulation software for the French gas distribution network (200,000 km of pipelines). National deployment in progress.

  • Replaced manual processes
  • Nonlinear optimization (Ipopt) and MILP (HiGHS)
  • Julia, JuMP

Atoptima — Supply chain SaaS

Lead Developer for a DeepTech startup (Inria spin-off). Last-mile delivery planning SaaS.

  • Full stack design and development
  • Julia, Rust, TypeScript, Angular, PostgreSQL, AWS
  • Recruitment and management

Research — Inria Realopt

PhD on exact methods for two-echelon vehicle routing problems (urban logistics). Realopt team was a world reference in combinatorial optimization.

  • Finalist of the VeRoLog Doctoral Dissertation Prize
  • 3 major publications
  • 160+ citations

Coluna.jl — Open source

R&D contribution to the Julia ecosystem. Branch-cut-and-price framework.

  • Lead developer
  • 200+ stars on GitHub

An optimization problem to solve?