DLeader, Inc.
DLeader Inc. develops KumiChem, an agentic AI platform that automates ASO and siRNA design for RNA therapeutics discovery. The platform uses specialized AI agents to handle target identification, binding prediction, structure folding, and chemical modification, generating ranked lead candidates up to 200× faster than manual workflows. KumiChem serves pharmaceutical companies, academic labs, and clinical research teams across rare disease, cardiometabolic, and neurology therapeutic areas.
- Drug Discovery & Therapeutics
- Artificial Intelligence
- AI Agents
- Biotechnology
- Software Only
Funding
Funding not disclosed
Founders
Product
Problem
RNA therapeutics discovery is a complex, multi-step process requiring researchers to integrate fragmented scientific data, specialized computational models, and experimental workflows. This complexity slows the path from target identification to validated drug candidates, particularly in therapeutic areas like rare disease, cardiometabolic conditions, and neurology where effective treatments remain limited.
Solution
KumiChem provides an AI-driven platform that automates the RNA drug discovery workflow from target identification through candidate validation. Users start with a simple text prompt, such as "Design siRNA for TTR mRNA," and the platform deploys specialized AI agents that handle splicing and accessibility analysis, binding prediction, structure folding, and chemical modification design. The system generates 24 ranked lead candidates evaluated for efficacy, safety, and IP clearance, with the top 5 candidates ready for real-world in vitro validation. The platform integrates a MetaAgent architecture that batch-processes unlimited items and synthesizes summaries to maintain awareness of all content without context constraints.
Target Audience
Primary users are pharmaceutical teams, principal investigators and university labs, and learners in the drug discovery field who need AI tools to accelerate RNA therapeutics research and development.
Features
- Target Identification Agent that identifies splice sites, exon junctions, and accessible mRNA regions, including cryptic splice variant mapping
- Binding Prediction Agent for efficacy and off-target screening
- Structure Folding Agent that predicts siRNA duplex stability and guide strand selection based on 5′ terminal thermodynamic asymmetry
- Optimization Agent applying 2′-OMe, 2′-F, and phosphorothioate modifications, plus GalNAc conjugation design for hepatocyte delivery
- Output of 24 lead candidates ranked by efficacy, safety, and IP clearance, with 5 candidates validated through real-world assays
- MetaAgent architecture providing infinite memory through batch processing and summary synthesis
- Benchmark-evaluated agents capable of identifying FDA-approved drug candidates (nucleic acid sequences, proteins, or small molecules) and discovering novel drugs
- Proprietary toxicity and modification datapoints exceeding 100K, supporting in silico safety screening