Fast-Tracking Formulations: The AI-Driven Future of Beauty and Pharma

Two University of Miami researchers have developed a patent-pending algorithm that could revolutionize the creation of products ranging from cosmetics and pharmaceuticals to coatings and food.
Fast-Tracking Formulations: The AI-Driven Future of Beauty and Pharma
College of Engineering Professor of Practice Samiul Amin and College of Arts and Sciences Professor Yelena Yesha. (Photo by Betsy Martinez/University of Miami)

When a customer buys a bottle of shampoo, they expect it to be of a certain consistency, foam up, and clean their hair. The expectations of what a product should do start with the perfect combination of what is sometimes more than a dozen ingredients listed on the back of a bottle. 

While the formulations may be hard to understand for some consumers, chemical engineering professor of practice Samiul Amin at the College of Engineering works to create the ideal combinations of ingredients spanning multiple industries that create formulated products. 

"It's a difficult task because there are a lot of different combinations of ingredients and many brands trying to compete on similar products," said Amin. "So, companies try to create small variations to stand out, like adding an anti-aging component to a moisturizing lotion." 

Products, including food items, rely on such carefully crafted formulas that even a small change can disrupt the entire product. Any type of modification can require extensive and labor-intensive research and development (R&D), yet consumer trends and needs evolve quickly. For example, plant-based proteins and alternative sweeteners are recent trends driven by consumer demand for healthier, more sustainable food options.  

"Traditional R&D is not going to cut it anymore," said Amin. "If you look across the consumer industry, there's a very big drive for sustainability. There's also a big drive for customization and personalization." 

That's where artificial intelligence comes in. 

A unique collaboration 

While researching how to speed up the traditional trial and error process, Amin met Yelena Yesha, a computer science professor at the College of Arts and Sciences and Knight Foundation endowed chair at the Frost Institute for Data Science and Computing.

While Amin focused on developing formulations, Yesha integrated that data into machine learning algorithms. Together, they discovered that artificial intelligence could help predict how certain ingredients in formulations interact and behave. They began training AI models with new data that allowed for more accurate predictions—eliminating time-consuming and costly experiments. 

"This collaboration is phenomenal because it merges deep expertise in AI with the specialized knowledge of chemical engineering and materials science," said Yesha. "By combining these two worlds, we can disrupt not just the chemical industry, but various sectors like agriculture and pharma." 

Fast-tracking the formulation of products 

The success of fusing chemical engineering and artificial intelligence led Amin and Yesha to launch Fast Formulator. Working closely with U Innovation, they licensed the intellectual property through the university and spun out the company. The startup aims to replace the process of chemical formulation and product development that relies on traditional experiments for trial and error with a predictive approach that uses artificial intelligence to quickly create new product formulations.  

With five formulators and five AI scientists on board, Amin and Yesha plan to license the software they’re creating to companies developing chemicals, cosmetics, cleaning products, biopharma, foods, inks, and more. The patent-pending algorithm has already piqued the interest of several multinational companies due to its potential to reduce R&D time and costs by as much as 75%. Their first product release is set for January. 

Revolutionizing pharmaceuticals and beyond 

This breakthrough technology also has the potential to revolutionize the pharmaceutical industry. Many monoclonal antibodies used for treating illnesses like rheumatoid arthritis, cancer, and diabetes are typically liquid injectables that face challenges such as stability and high viscosity. The AI models being developed could help predict the right conditions and ingredients to ensure the drugs remain safe and effective for patients. 

Amin and Yesha are not just focused on technology; they are also committed to education. 

"We want to train the next generation of engineers and scientists to be equipped for these emerging technologies," said Amin.  

__________________________________________________________________________________ 

The College of Engineering will host the 2nd Annual Rothberg Catalyzer AI Summit at the Shalala Center on February 11, 2025.  



Top