Inventum.ai

Inventum.AI is an AI-driven drug discovery platform that combines generative artificial intelligence with experimental validation to accelerate the design of novel, effective, and safe therapeutics. The platform serves pharmaceutical researchers, medicinal chemists, and drug discovery organizations seeking to reduce costs and timelines at the early stages of R&D. By integrating cutting-edge AI with in vitro validation, Inventum.AI delivers precise affinity and ADMET predictions within an intuitive, user-friendly interface — enabling users to design molecules in just four clicks and generate structures in 72 hours instead of weeks.

The platform has demonstrated measurable impact, including a 40% cost reduction at early R&D stages, a fivefold acceleration of research and development processes, and a 25% increase in prediction accuracy. Inventum.AI's mission is to generate targeted molecular structures tailored to medicinal chemists' needs, democratizing drug discovery and empowering a broader range of organizations to contribute to innovative treatment development.

Core Platform Capabilities

• Ligand Generation: Molecular structures are constructed atom by atom using the Nature Based Generator (NBG), which selects optimal atom types by analyzing the macromolecular context through deep learning models trained on high-resolution crystallographic data.
• Binding Site Identification: The proprietary AI-based binding site search model, SiteRadar, has demonstrated superior performance over existing solutions in terms of binding site prediction accuracy.
• Molecular Docking and Interaction Analysis: The platform covers molecular docking, scaffold growth, and detailed interaction analysis to ensure strong complementarity between generated structures and the ligand-protein binding site.
• ADMET Predictions: Precise predictions of absorption, distribution, metabolism, excretion, and toxicity properties are delivered to support drug-likeness and safety assessments.
• Step-by-Step Generation Process: The generation workflow involves positioning a selected starting fragment, generating the scaffold, and generating the periphery — with the ability to select chemotypes of interest at each stage, directing generation toward the most relevant chemical space.

Key Differentiators and Scientific Approach

• Nature-Inspired Methodology: The platform's generative AI is trained on empirically derived data from existing protein-ligand complexes using high-resolution crystallographic structures, faithfully reproducing molecular recognition principles observed in natural systems and reducing dependency on artificial parameters.
• Novelty: Models construct molecular structures atom by atom, enabling access to a diverse and expansive chemical space.
• High Specificity: Optimized atom types within macromolecular environments foster the generation of structures with strong binding site complementarity.
• Drug-likeness: Proprietary criteria are used to evaluate generated structures against up-to-date medicinal chemistry standards, ensuring only drug-like and synthesizable compounds are produced.
• Experimental Validation: All generated molecules are rigorously tested in vitro to confirm efficacy and safety, with the NBG algorithm having successfully reproduced the crystal structure of a CDK2 inhibitor and demonstrating strong correlation with the established Glide (Schrödinger) scoring function.

Therapeutic Focus and Case Studies

• Inventum.AI's pipeline targets oncology, metabolic disorders, and rare diseases, leveraging advanced AI-driven methodologies to identify novel therapeutic candidates and enhance the success rate of drug development.
• Published case studies include an AI-driven discovery of an IRAK4 inhibitor (December 2024) and the design and validation of FLT3 inhibitors for Acute Myeloid Leukemia (AML) treatment (January 2025).

Leadership and Team

• Ramil Kuleev (CEO, co-founder) holds a PhD from Kazan State University and previously founded and led the AI Institute at Innopolis University, overseeing 50+ projects with revenues exceeding $20 million. He also developed an AI radiology service for chest X-rays in collaboration with Moscow's Health Department and founded Pulmoscreen, a radiology software startup deployed in over 10 hospitals.
• Ruslan Lukin (CSO, co-founder) is a PhD Candidate in Physical Chemistry and previously served as Principal Investigator at the AI in Materials Lab at Innopolis University, leading international collaborations and achieving 2nd place in the Open Catalyst 2021 Challenge by Meta AI. He has authored over 15 publications including papers in Nature Communications and holds 5 patents.
• Sergei Evteev (Computational and Medicinal Chemistry Lead) holds a PharmD from Moscow State University and brings experience from Insilico Medicine, where he conducted computer-aided drug design (CADD) for various therapeutic targets and contributed to software development projects including Chemistry42 and Generative Biologics.
• The advisory board includes Dr. Alexander Tropsha, K.H. Lee Distinguished Professor and Associate Dean for Pharmacoinformatics and Data Science at UNC Eshelman School of Pharmacy.

Inventum.AI brings together expertise in AI, computational chemistry, and structural biology to translate computational insights into real-world therapeutic applications, offering a platform that is both scientifically rigorous and accessible to a wide range of drug discovery stakeholders.