Our mission
To become the intelligent decision engine behind every complex operation—where planners steer, and algorithms empower.
Why "Penumbra"
In every optimization problem, there’s a boundary:
A line between what’s feasible and what’s infeasible.
Classic models chase the optimal point—balanced precisely on that line. But real-world operations aren’t clean. They’re noisy, dynamic, full of hidden constraints and shifting inputs. What looks optimal in theory often breaks in practice.
We live in the penumbra—the region of partial shadow between light and dark, between rigid feasibility and outright failure.
That’s where real decisions happen.
Where flexibility matters more than mathematical purity.
Where a solution that works well enough and adapts fast enough outperforms one that’s theoretically perfect but fragile.
Penumbra Optimization exists to solve in that region—where robustness, adaptability, and human judgment meet advanced algorithms.
Because in operations, the real edge isn’t a line. It’s a zone.
Experienced Team

Reginald Dewil (Co-Founder)
Reginald has over 15 years of experience tackling decision problems in process planning, routing, scheduling, and planning in highly complex dynamic problem settings. He holds a PhD in Engineering Science, specializing in Operations Research.
Reginald is passionate about delivering solutions that truly integrate into clients’ operations, driving measurable improvements in efficiency and throughput. He wrangles the AI pipeline and oversees product direction, bridging the gap between cutting-edge optimization technology and real-world scheduling needs.
Carolien Lavigne (Co-Founder)
Carolien brings deep expertise in operational research and advanced planning algorithms. With a PhD in Applied Economics, she has spent over a decade working on optimal decision-making in logistics, waste management, and manufacturing.
Carolien works on the core optimization components of the AI system, both on the problem-agnostic metaheuristic components and the highly specialized problem-specific decision rules and local search operators. She ensures that the AI engine is both powerful and efficient.