“Traditional methods for companies in the chemical industry have been mostly trial-and-error based for their product development. Implying they take a batch of chemicals, mix them together and keep on repeating experiments after experiments until they find the desired properties. On average a company spends multiple batches of experiments to find the best set of ingredients,” explains Kunal Sandeep, co-founder of Polymerize.
The Singapore based startup Polymerize, which is part of the 5-HT ecosystem is working on a solution for this problem: an AI Software for Polymer and materials companies to expedite their product development timeline. In an interview with 5-HT, Kunal Sandeep explains the advantages of their software platform, continuous production processes and modularised production.
“Our proprietary machine learning algorithms can deliver highly accurate formulation recommendations, reducing the number of experiments you need to run to find the perfect formulation. In practicality this means we can help find the right set of ingredients in 15 to 20 experiments cutting down time and cost for companies. We make sure the information our clients need is available right at the tip of their fingers. We are digitalizing the entire workflow automatically. We believe we can actually help companies to get their products to the market three times faster by using our software platform.”
For whom Polymerize is intended
“Polymerize’s software platform provides materials and polymer companies with access to faster formulation development using our proprietary AI technologies, while enabling smooth data management leading to materials innovation and increased operational efficiency in their R&D activities.
The market needs a product like ours to expedite product development and get to the market faster to meet the consumer demand curve. For example, with rapidly growing demands in electric vehicles, healthcare, biodegradable products and many more, there is a lot of pressure coming onto the materials and chemical companies to find and develop the right materials faster than before. And this cannot be done using the age-old traditional practices where companies had to spend long cycles of experimentation to get the right set of formulations/ingredients for the desired set of properties. We provide the required software and AI technology to help companies accelerate product development in 1/3 of the conventional time and get to the market much faster.”
Using a domain guided Machine Learning approach
“Ours is a fairly new company, we just started in May of 2020 – the fundamental difference we have seen compared to other existing competitors is that we use a domain guided machine learning approach unlike our competitors. We work closely with our customers and use our proprietary algorithms that infuse machine learning with domain knowledge, unlike the competitors. The domain knowledge is centered around the materials, processing, properties and the R&D specialist’s experience. Our software gives the computing edge to the R&D scientists to search materials faster, run predictions, collaborate with other scientists and plan experiments better. We help them get rid of traditional spreadsheet-based formulation development and bring operational efficiency.“
Understanding the value of cloud technology due to COVID-19
“The pandemic has had a very positive impact on the nature of our business. Pre COVID-19 everybody used to go to the labs. Everybody used to do their own experiments, look at the boards, talk to people and so on. Suddenly we are dealing with a global pandemic. People are bounded by their own physical setting and rely on technology to access information remotely. A platform like ours is going to help them a lot. Because now people can access all the reports and predict things without going to the factory. Their product development is not getting backlogged, instead the process is getting smoother now. I feel like COVID-19 is actually helping to understand the value of cloud technology and remote working.”
From startup to standard
As a company we want to lead the forefront of material innovation for the materials and chemical companies. For traditional data science models, you need a lot of data points and for new formulations companies have relatively fewer data points. To cater to the few data points, we need to supplement them with domain information to make AI-based predictions more accurate. We are creating the much needed data infrastructure for chemical companies across South-East-Asia, Asia and Germany to be AI-ready. It is important to us to build our base very strongly so that companies can do things faster and much better. We want to ensure that the technology becomes an integral part of their processes. And not just an external information system that we need to integrate.
We have a fast onboarding process. We start with a three-month trial period. Clients give us their objectives to test our capabilities. Once we give them satisfactory results, we move to collaboration. At the moment, we have a subscription model and engage with customers on an enterprise-wide license agreement . “
The timing is right to leap ahead into a world of technology
“We are consistently experimenting towards perfecting our AI capabilities. We are hopeful to contribute to the open-source world and invite developer communities and scientists to build on our technology that we are currently working on that can eventually bring in the digital enablement for the chemical companies. Opportunities are among companies like ours.”