Master the complete formulation optimization pipeline. From DoE design to Bayesian lab-in-the-loop optimization, ICH Q8 design space, and publication-grade reports - no programming required.
From DoE design fundamentals to Bayesian lab-in-the-loop optimization - a complete formulation science toolkit
Design experiments using Box-Behnken, CCD, D-optimal, factorial, Latin hypercube, and mixture designs
Fit quadratic OLS response surface models, interpret effect plots, and assess lack-of-fit
Compare Gaussian Process, gradient boosting, and RSM using nested cross-validation on your own data
Map Pr(specs met) heatmaps aligned with ICH Q8 / QbD quality-by-design requirements
Run NSGA-II, visualize trade-offs with parallel coordinates and scatter matrices
Use Bayesian optimization to suggest the next experiment on the desirability scale - close the experimental loop
Apply Monte Carlo perturbation around candidates and define sympy constraint expressions
Export PDF reports, CSV bundles, and 26 figure types ready for journal submission
This workshop is for anyone working with experimental formulation data and wanting to go beyond basic statistical analysis - no programming needed. Everything is point-and-click.
Pharmaceutical R&D scientists who run DoE experiments and want smarter optimization, probabilistic design space maps, and efficient next-experiment suggestions.
Graduate and PhD students working on formulation theses or projects. Learn to analyze your experimental data and generate journal-quality results without writing code.
Regulatory and process development scientists who need ICH Q8 probabilistic design spaces and must demonstrate that their formulation process is robust under variability.
Manufacturing and process engineers who want to optimize multi-variable processes, visualize trade-offs across competing objectives, and identify robust operating regions.
Prerequisites: Experience running lab experiments, comfort with Excel, and basic understanding of means and standard deviations. No Python, R, or ML background required.
A complete formulation optimization environment - from raw DoE data to ICH Q8 design space and publication-ready reports
Generate statistically efficient experiment plans for any factor space
Full factorial, fractional factorial, and Plackett-Burman designs for factor screening
Quadratic designs for fitting curved response surfaces to find optima
Latin hypercube sampling and D-optimal designs for irregular factor spaces
Simplex lattice and simplex centroid designs where components must sum to a constant
RSM polynomial regression and modern ML - compared honestly side-by-side
Classical quadratic and cubic OLS models with effect plots, lack-of-fit, and ANOVA breakdown
Bayesian non-parametric model with built-in uncertainty estimates - ideal for BO acquisition
Ensemble tree models with nested CV - often outperforms RSM on nonlinear, noisy data
Everything you need to find the optimum, map the design space, and report results
Lab-in-the-loop next-experiment suggestions using Expected Improvement on the desirability scale
NSGA-II evolutionary optimization for competing responses with interactive trade-off visualization
Pr(specs met) heatmaps across the factor space - ICH Q8 compliant design space definition
Monte Carlo perturbation around a candidate point to verify stability against real-world variability
Auto-generated plots at journal resolution - from contour maps to 3D response surfaces
Complete project report and raw data bundle ready for inclusion in a journal submission or regulatory dossier
Date TBA · 09:00 - 15:00 · 1-hour lunch break 12:00 - 13:00
| Time | Session | Topic |
|---|---|---|
| 09:00 - 09:15 | Welcome | Orientation, credentials, sample dataset download |
| 09:15 - 10:00 | Session 1 | DoE Fundamentals - designs, factor types, response selection, randomization |
| 10:00 - 10:45 | Session 2 | RSM & Effect Modeling - polynomial OLS, main effects, interactions, ANOVA |
| 10:45 - 11:00 | Break | |
| 11:00 - 11:45 | Session 3 | ML vs RSM + Bayesian Optimization - Gaussian Process, gradient boosting, Expected Improvement |
| 11:45 - 12:00 | Lab 1 | Hands-on: Upload dataset, generate DoE, fit RSM model (tablet dataset) |
| 12:00 - 13:00 | Lunch | |
| 13:00 - 13:45 | Lab 2 | Hands-on: ML comparison + BO - which model fits best? Next-experiment suggestion |
| 13:45 - 14:30 | Session 4 | Probabilistic Design Space & Pareto Optimization - ICH Q8 maps, NSGA-II trade-offs |
| 14:30 - 14:50 | Lab 3 | Hands-on: Full pipeline on your own dataset - design space map + PDF report export |
| 14:50 - 15:00 | Closing | Key takeaways, next steps, certificates of completion |
Includes 30-day full access to the InsilicoΣ FORMULA-X platform after the workshop
Per participant
Per participant
Complete the form below. We'll send confirmation and payment details to your email.
Chrome, Firefox, or Edge. The platform is fully browser-based - no software installations needed.
You should have run experiments before and understand what factors and responses mean in your domain.
Comfortable with spreadsheets and basic concepts like mean, standard deviation, and p-values. No Python or R required.