BenchSci

BenchSci is an AI-driven life sciences company on a mission to unravel the complexity of disease biology for the betterment of patients. Founded in 2017 and headquartered at the intersection of machine learning and biomedical research, BenchSci builds neurosymbolic AI solutions that help preclinical R&D teams move from hypothesis to successful experiment in days rather than years. The company serves thousands of scientists across leading academic institutions and major pharmaceutical organizations, addressing the root cause of drug discovery failure: a fundamental lack of understanding of underlying disease biology.

At the heart of BenchSci's offering is ASCEND, the world's first neurosymbolic AI platform designed to transform how biopharma organizations understand disease biology. ASCEND acts as a co-pilot for preclinical teams, guiding ideation, experimental design, and hypothesis generation to accelerate breakthroughs and bring effective therapies to patients faster. The platform is powered by a rigorously vetted Biological Evidence Knowledge Graph (BEKG), curated by industry experts to ensure all information is backed by traceable experimental evidence.

ASCEND Platform and Core Technology

• ASCEND is the first neurosymbolic AI platform built specifically to decode disease biology, combining a structured knowledge graph with advanced foundation models to deliver AI copilots and co-scientists for drug discovery teams.
• The Biological Evidence Knowledge Graph (BEKG) structures over 400 million entities and more than one billion relationships, with every data point linked directly to traceable experimental sources.
• The platform integrates diverse scientific data types including peer-reviewed publications, multi-omics datasets, and clinical trial data to provide a unified, comprehensive view of diseases.
• A proprietary multimodal AI extracts insights from text, experimental figures, and supplementary files, enabling deep understanding of published and internal research.
• BenchSci can ingest a client's proprietary internal data — such as lab notebooks — to create a customized, secure map of disease biology tailored to each organization's needs.
• ASCEND supports natural language queries that are translated into complex, multi-hop scientific reasoning, enabling a deeper level of discovery that mirrors how scientists think.

Key Differentiators

• Unparalleled data access: Unlike platforms limited to public data, BenchSci's knowledge graph integrates tens of millions of scientific publications, including closed-access journals through direct agreements with major publishers, alongside licensed, public, and client-proprietary data.
• Data quality and AI hallucination mitigation: The BEKG serves as an evidence backbone ensuring scientific validity. Every relationship and entity in the graph is linked to experimental sources, reducing the risk of AI-generated inaccuracies.
• Human-in-the-loop curation: A dedicated team of over one hundred scientists provides ongoing curation of the knowledge graph, ensuring scientific consistency and differentiating BenchSci from generic AI models.
• Deep biopharma workflow integration: The platform is designed not as a standalone tool but as an integrated solution that supports and enhances existing scientific workflows, augmenting rather than replacing researchers.
• Dedicated deployment and change management support: Each client account receives a dedicated team including PhD-level scientific liaisons who work directly with researchers to integrate the platform into existing workflows and maximize adoption.

Company History and Milestones

• BenchSci was founded in 2015 when co-founders Tom Leung, David Q. Chen, Elvis Wianda, and Liran Belenzon met at the University of Toronto's Creative Destruction Lab and developed an AI solution to the antibody reproducibility crisis.
• In 2017, BenchSci launched its first commercial application — AI-Assisted Antibody Selection — using experiment-specific text machine learning and proprietary vision ML models to help scientists reduce experimental failure.
• The company subsequently expanded its technology to guide the selection of many types of reagents beyond antibodies.
• Within three years of launch, more than 4,300 leading academic research institutions and 16 of the top 20 pharmaceutical companies were using BenchSci's AI solutions, with over 50,000 scientists relying on the platform.
• BenchSci launched Project Butterfly — developed in collaboration with top pharma R&D leaders — to apply its visual ML capabilities to build the first objective, evidence-backed map of underlying disease biology.
• The portfolio success platform developed during Project Butterfly was named ASCEND, designed to increase the speed and yield of R&D portfolios by multiples, not just percentages.
• BenchSci completed a Series D funding round led by Generation Investment Management, enabling further enhancement of ASCEND, expansion of offerings, and continued support for scientists worldwide.
• The company has integrated generative AI into its platform in an ethical and responsible manner, spanning data extraction, contextualization, summarization, and conversational AI features.

Strategic Partnerships and Customer Base

• BenchSci has established a strategic partnership with Thermo Fisher Scientific to bring its AI capabilities to more researchers globally.
• The company has signed new contracts with Sanofi and Merck to accelerate drug discovery, enhance research efficiency, and drive innovation through AI and machine learning.
• BenchSci's advisory board includes senior leaders from Pfizer, Sanofi, Amgen, Bayer Pharmaceuticals, Johnson & Johnson, Takeda Pharmaceuticals, and Roche Pharmaceutical, as well as experts in machine learning and venture capital.

Leadership and Team

• BenchSci's leadership team brings expertise spanning pharmaceutical R&D, machine learning, data science, and business operations, united by a commitment to transforming how life-saving medicines are discovered.
• The founding team includes CEO Liran Belenzon, Chief Science Officer Tom Leung, and Chief Data Officer Elvis Wianda, Ph.D., alongside a broad executive team covering technology, commercial, finance, talent, and customer success functions.
• The company employs scientists, engineers, product managers, and business leaders all focused on applying AI to improve human health.

BenchSci's neurosymbolic AI approach — combining a rigorously curated biological evidence knowledge graph with advanced AI models and deep scientific expertise — positions the company as a world leader in AI solutions for preclinical R&D, with the goal of dramatically accelerating the delivery of effective therapies to patients who need them most.