ICARO (Intelligent Comparative Assessment and Research Observatory) is a cutting-edge research lab focused on advancing the field of large language model evaluation. We are part of DEXAI, a leading European AI research organization dedicated to developing safer, more reliable, and more transparent artificial intelligence systems.

Our mission is to create robust evaluation frameworks that can accurately assess the capabilities, limitations, and potential risks of modern AI systems. Through rigorous benchmarking, novel assessment methodologies, and open research, we contribute to building trust in AI technologies.

Model Evaluation Research

We develop comprehensive evaluation protocols for assessing language models across various dimensions including reasoning, factual accuracy, safety, robustness, and ethical considerations. Our research helps identify model strengths and weaknesses across diverse tasks and domains.

Benchmark Development

We create and curate high-quality benchmarks that reflect real-world challenges and requirements. Our benchmarks are designed to be fair, comprehensive, and continuously evolving to keep pace with rapid developments in AI capabilities.

Open Science Initiative

We believe in transparent, reproducible research. All our evaluation frameworks, datasets, and methodologies are made publicly available to promote scientific collaboration and accelerate progress in the field.

Industry Collaboration

We work closely with industry partners to ensure our evaluation methods are practical and applicable to real-world deployment scenarios. Our research informs best practices for AI system development and deployment.

Join Us

We are always looking for talented researchers, engineers, and students passionate about AI safety and evaluation. Whether you're interested in academic research, open-source development, or industry collaboration, we'd love to hear from you.

Open Positions

No open positions at this time.