Search papers, labs, and topics across Lattice.
CURATE is a novel human-in-the-loop multi-agent system that leverages LLMs to manage the entire lifecycle of composable workflows, addressing the gap in existing coding agents that focus solely on code generation. By incorporating a module catalog for the storage and reuse of components, CURATE facilitates the sharing of curated modules and subgraphs, enhancing reproducibility and collaboration in software development. The system's feasibility was validated through experiments that reproduced and adapted workflows from the SeBS-Flow benchmark and automated a complex environmental engineering workflow, demonstrating significant potential for accelerating applied and scientific development.
CURATE transforms workflow management by enabling seamless composition, reuse, and deployment of code, bridging the gap between generation and execution.
Agentic code generation has shown promise in automating and accelerating software development by utilizing Large Language Models (LLMs) to generate, test, and deploy code. For engineers and scientists, such systems have the potential to accelerate the development of applied and scientific workflows while reducing barriers to entry in domains that have yet to fully realize their benefits. However, a key gap remains: existing coding agents primarily focus on code generation and do not address the entire workflow lifecycle, including deployment and sharing. As a result, users develop and stitch modules independently while managing deployment on their own. To address this gap, we propose CURATE - Composition, User-in-the-loop, Reuse, and Automated Task Execution - a novel human-in-the-loop multi-agent system that uses LLM agents to manage and develop composable workflows across their entire lifecycle. A key feature of the system is a catalog that allows for the storage and reuse of modules across workflows. Module catalogs provide a foundation that can be expanded to support FAIR principles by facilitating the sharing and reuse of curated modules and subgraphs. We demonstrate the feasibility of our system with an initial prototype using Claude Opus 4.8, comprising 6 experiments: reproducing and adapting 4 workflows derived from the SeBS-Flow benchmark suite, and automating the development and scaling of a workflow that leverages a complex mechanistic model in environmental engineering used to simulate anaerobic digestion.
