Meet Nike — Persist’s AI Formulation Scientist
Our proprietary AI formulation engine analyzes your drug development challenge and delivers data-driven formulation strategies within 24 hours.
Example Questions:
"I want an oral formulation for aspirin that avoids hydrolysis and improves stability at room temperature."
"I want a SEDDS formulation for a poorly soluble drug (log P of 4, MW 1,000) that achieves 20% drug loading while preventing crystallization during storage."
⚠️ Important: Do Not Enter Proprietary Information
-
Use non-proprietary SMILES codes only
-
Describe your drug in general terms (e.g., "poorly soluble small molecule, log P of 3")
-
Use generic drugs as examples (e.g., aspirin, ibuprofen, metformin)
What is Nike?
Nike is Persist AI's formulation intelligence engine — trained on thousands of formulation experiments and powered by advanced computational chemistry.
Formulation Composition
Optimal excipient selection & ratios for your target product profile
Long-Term Shelf Life
Predict instability and impurities early in development
In-Vivo Performance
PBPK-informed formulation design for better bioavailability
API CMC Risk
Identify API vulnerabilities to solve through formulation
Work With What You Have
📊 Molecular Descriptors
Don't have a full structure? No problem. Nike can work with molecular properties like log P, molecular weight, solubility class, and other descriptors.
🧬 Full Structure (SMILES)
Providing a complete molecular structure gives Nike more predictive power for excipient compatibility and stability predictions.
Nike’s Multi-Agent Intelligence Engine
Nike uses a multi-agent AI system where specialized agents work together to analyze your formulation question and provide comprehensive recommendations.
Global Orchestrator Agent
Formulation
Agent
Prediction
Agent
CMC Risk
Assessment Agent
Toxicology & Bioavailability Agent
Experimentation
Agent
AI & Computational Capabilities
Molecular Modeling
-
Density Functional Theory (DFT) for solubility prediction
-
Formulation stability assessment a priori
-
Molecular Dynamics for drug-polymer interactions
-
Drug-excipient compatibility screening
Predictive AI
-
Bayesian Optimization — find optimal formulations faster
-
Deep Neural Networks with transfer learning
-
Works even with limited API availability
-
Multi-Agent Workflows for de-risking
Process Modeling
-
Film & Wurster coating simulation
-
Lyophilization cycle optimization
-
Spray drying & hot melt extrusion
-
Tableting & microencapsulation
Case Studies
Unstable Phase 2 Formulation
Problem: Customer had an unstable Phase 2 formulation with limited design space explored
Solution: AI-driven analysis of hydrophobic pockets and excipient behavior found ideal excipients that improved stability
Problem: Customer struggled for 3 years testing 10-15 formulations/month with no viable candidate
Solution: 700 formulations with AI-driven insights in 3 months — lead formulation identified and moved to clinic
3 Years → 3 Months
Problem: Multinational customer had labs generating characterization data that didn't match
Solution: Cloud laboratory testing generated uniform data set for AI training — saved 40 months of effort
Global Data Harmonization
Powered by Real Data
Nike isn't just trained on literature — it's built on thousands of real formulation experiments from Persist AI's high-throughput robotic laboratory.
-
1,000+ formulations tested per month
-
Integrated HPLC, spectrophotometry, and Raman analytics
-
Continuous learning from new experimental data
