



fewer experiments than traditional DoE in small-molecule process optimization (Merck)
higher yields than DoE across 11 parameters in 42 experiments (Merck, Biochemistry 2024)
of the full factorial space sampled to reach an 80%+ yield hit (Boehringer Ingelheim, OPRD 2026)
reduction in enzyme loading in a biocatalytic cascade optimization (Merck)
Stamato Solutions & Meridion Technologies · Chimica Oggi – Chemistry Today, Vol. 43, 2025
Bayesian-optimization models built from over 80 experimental runs to characterize a spray freeze-drying process.
SuntheticsML doesn't replace scientific judgment. It makes every experiment you run smarter, so the decisions you make next are better supported.
SuntheticsML is designed to work with limited experimental data — you can start a campaign with as few as 5 to 10 data points. Unlike traditional machine learning tools that require hundreds of experiments, SuntheticsML uses advanced Bayesian optimization and active learning algorithms purpose-built for chemistry R&D. The platform intelligently suggests the most informative next experiments to run, so you accumulate insight faster with fewer runs. As your dataset grows, the model’s predictions become more accurate, but you never need to front-load a large batch of experiments before seeing value. Whether you’re starting a new synthesis route or optimizing an existing pharmaceutical process, SuntheticsML meets you where your data is.
SuntheticsML does not replace Design of Experiments (DOE) — it enhances it. Traditional DOE methods like full factorial and response surface designs are powerful, but they require large, predefined experimental grids that can be expensive and time-consuming in pharmaceutical R&D. SuntheticsML uses Bayesian optimization to dynamically select the next best experiment based on what the model has already learned, making it far more data-efficient than static DOE designs. The result is a smarter, iterative campaign that converges on optimal conditions faster. Scientists who are comfortable with DOE will find SuntheticsML familiar in goal — understand the chemical space and find the optimum — but far more efficient in execution.
No coding or machine learning expertise is required to use SuntheticsML. The platform is built for bench chemists and process scientists, not data engineers. Your team logs experimental conditions and outcomes through a structured interface, and SuntheticsML handles the modeling, optimization, and next-experiment recommendations automatically. There are no scripts to write, no models to train, and no hyperparameters to tune. Sunthetics’ algorithms are purpose-built for small-data chemistry environments, so the platform works out of the box for pharmaceutical synthesis, formulation, and process optimization workflows without any AI or programming background.
Your scientists do. SuntheticsML recommends the next experiment to run based on its predictive model, but the final decision always rests with your team. The platform surfaces ranked suggestions with confidence intervals and predicted outcomes, giving chemists full transparency into why a particular set of conditions is being proposed. Scientists can accept a recommendation, modify parameters based on practical constraints like reagent availability or safety considerations, or override the model entirely. This human-in-the-loop approach is fundamental to how Sunthetics works — the AI accelerates decision-making, but domain expertise and scientific judgment remain in control throughout the campaign.
SuntheticsML supports integration with automated lab platforms and high-throughput experimentation workflows. Data from automated reactors, flow chemistry systems, and analytical instruments can be fed into the platform to close the optimization loop without manual data entry. For pharmaceutical teams running high-throughput screening or automated synthesis campaigns, this means SuntheticsML can operate as a continuous optimization engine — receiving results, updating its model, and generating the next set of experimental conditions in near real time. Contact the Sunthetics team to discuss your specific instrument stack and integration requirements.