Convexia uses specialized AI agents and expert human review to source and evaluate overlooked drug assets globally. The platform conducts deep in silico simulations assessing binding, ADME/PK, and toxicity to predict clinical feasibility and operational risk. This integrated approach provides clients with high-probability-of-success candidates benchmarked against market and historical data.
Funding
Funding not disclosed
Founders
Product
Problem
Identifying and evaluating high-potential drug assets from vast global datasets is a time-consuming and resource-intensive process for pharmaceutical companies. Traditional due diligence workflows often lack the depth and speed required to accurately assess scientific validity, market viability, and operational risks for preclinical and IND-stage candidates.
Solution
Convexia provides an AI-driven platform designed to accelerate and de-risk early-stage drug asset evaluation for pharmaceutical partners. Our system systematically scans global data sources to identify overlooked drug candidates, performing deep in silico simulations to assess scientific merit. A human-in-the-loop approach integrates expert review at critical decision points, ensuring robust validation of biological rationale, market potential, and operational feasibility. This integrated methodology significantly reduces asset evaluation timelines, enabling partners to make more informed investment decisions and advance promising therapeutics more efficiently.
Target Audience
Our primary customers are pharmaceutical companies and biotech firms seeking to enhance their early-stage drug discovery and business development pipelines by leveraging advanced AI for asset evaluation.
Features
- AI-powered asset discovery agent that scans global databases, patent filings, and preclinical literature for IND-stage candidates.
- In silico simulation suite utilizing over 50 custom-tuned models for binding affinity, ADME/PK, immunogenicity, and mechanistic fit assessment.
- Specialized AI agents for evaluating market attractiveness, including unmet need, disease burden, IP protection, and payer risk.
- Operational risk assessment agent that models CRO fragility, CMC complexity, and site readiness to predict trial execution risk.
- Human-in-the-loop validation process with domain-specific PhDs and clinicians reviewing scientific, commercial, and operational data.
- Protein and sequence analysis leveraging models like ESM-3, ProtBERT, and AlphaFold for structure and ligandability prediction.
- Docking and structure analysis using tools such as DiffDock and GNINA to model binding poses and affinity.
- Consensus ADMET prediction across multiple models to identify potential toxicity liabilities early in the development cycle.