Independent research · Distributed

Independent minds. Shared standards. Research for the agentic era.

REQS Labs brings together independent researchers working on trusted autonomous systems, algorithms, and decision-making under uncertainty.

Selected work by members

Research in focus

2026 · Nature Computational Science
Large language models as human proxies
A review of four ways LLMs stand in for people, showing why each use needs its own criteria for validity.
Nikita Karetnikov, Iyad Rahwan, Davor Svetinovic
Member research · Trusted Agentic Systems Lab
2025 · Discrete Applied Mathematics
Dual bounded generation: Polynomial, second-order cone and positive semidefinite matrix inequalities
Establishes conditions for bounding families of infeasible solutions in nonlinear integer systems, supporting algorithms that enumerate feasible solutions.
Khaled Elbassioni
Member research background · Theory Lab

Selected publications retain their original authorship and institutional context.

Member labs

Led by Davor Svetinovic

Member lab

How can autonomous systems be trusted when they act across code, markets, and physical infrastructure?

Agent assuranceDecentralized systemsCyber-physical resilienceRequirements engineering

Led by Khaled Elbassioni

Member lab

Algorithms with guarantees for decisions at scale.

Algorithms and complexityOptimization under uncertaintyGames, incentives, and multi-agent decisionsTheory for high-impact systems

Davor Svetinovic presented planned research on agentic AI and autonomous finance at the ADIA Lab Symposium in Abu Dhabi on 28 October 2026.

Collaborate

Discuss joint research, a practical setting for evaluation, or a future member lab.