Quantum-inspired optimization for problems too large for classical solvers.

Icosa Computing builds optimization software, deployable on quantum and physics-enhanced hardware, that solves large combinatorial optimization problems for financial institutions (portfolio optimization, index tracking, pairs trading, currency arbitrage) at a scale conventional solvers can't handle.

Icosa Computing was founded in 2022 by Mert Esencan, who holds a BS in Physics and Symbolic Systems and an MS in Management Science and Engineering from Stanford, and a PhD in Physics from Oxford, where his thesis focuses on error-corrected quantum computation using time crystalline phases. He started the company after using annealing-based, quantum-inspired techniques to trade his own account, then proved the approach through pilots with financial institutions before building a product around it.

The next frontier in AI is not just bigger models, but more reliable, domain-aware models that hallucinate less and work inside real enterprise context. Mert's quantum-inspired approach targets exactly that problem: helping teams adapt models to their own documents, code, and terminology, while keeping everything local and private. With Zeno and LMShop, the company is turning a deep technical insight into a practical workflow for custom, on-device AI.