AI research and product company developing models that can work, communicate, and coordinate over long horizons
AI research and product company developing models that can work, communicate, and coordinate over long horizons. It researches reinforcement learning, multi-agent systems, and environment design, and offers OpenReward, a platform for serving environments to train and evaluate language models, plus BackSearch, a point-in-time web search and fetch API for reproducible forecasting, evaluation, and training.