The competition within the AI for Science sector is progressively shifting towards closed-loop scientific research. Recently, Levin BioTech, an AI-driven protein design company based in Hangzhou, introduced a groundbreaking research tool—Lévin™ Harness—designed to empower the research community. This innovative tool consolidates data, models, plug-in tools, computational resources, and workflows into a cohesive agent workspace. By harnessing the power of general large models, it comprehends user requirements, formulates action plans, and invokes specialized scientific models and analytical tools to execute tasks seamlessly. Researchers can effortlessly conduct studies by articulating their needs in natural language. Moreover, users have the flexibility to enhance the tool's capabilities by integrating new models through plug-ins and transforming successful research methodologies into reusable Workflows. This approach effectively tackles the challenge of maintaining sustained AI operations in research environments and significantly accelerates the journey from concept to realization.

Established in 2021, Levin BioTech derives its name from the Levinthal paradox, reflecting its commitment to pushing the boundaries of scientific discovery. The company's founding team boasts an interdisciplinary background and has independently developed Pallatom, a cutting-edge full-atom protein generation model. Recognized by NVIDIA as a representative foundational protein model, Pallatom has spawned a comprehensive product family. Concurrently, Levin BioTech has assembled algorithm and wet-lab experimental teams, fostering a unique 'umbrella structure' that has yielded numerous successful outcomes, including the development of steviol glycoside enzyme preparations and brazzein.

Lévin™ Harness originated from Levin BioTech's internal R&D efforts. In light of the existing technological infrastructure at research institutions, the tool offers unparalleled flexibility in model selection, plug-in expansion, and computational resource access. User data remains securely stored on their local machines, granting them full control over backup strategies. Furthermore, users can save reusable executable components as Workflows, enabling them to construct a proprietary AI research system that fosters the growth of an open ecosystem.

Levin BioTech's ultimate vision is to establish a '24-hour intelligent R&D factory.' Amidst the burgeoning array of point capabilities in the current AI for Science landscape, connecting these disparate elements has emerged as a formidable challenge. Levin BioTech is positioning itself as a key player in addressing these emerging issues, driving innovation and collaboration within the scientific community.