Ai-guided valorization of pectin side streams into functional materials
An AI-native materials-informatics platform combined with a lab-scale prototyping network to turn fibrous extracted-peel residue from pectin processing into high-value functional materials. The workflow screens and ranks valorization routes—fibre/polysaccharide fractions, bio-based composite feedstocks, and engineered adsorbents or biochar—while accounting for feedstock variability, low pH and residual nitrate. Active-learning design of experiments and predictive models link feedstock characteristics and process conditions to product properties and yield, enabling rapid, validated, scalable route selection.
Overview
Polymerize's AI-guided valorization solution applies a materials-informatics platform to pectin side streams, specifically the fibrous extracted-peel residue generated during pectin processing, and converts them into functional materials. Instead of treating feedstock variability as a drawback, the platform treats it as an input and systematically searches the product and process design space to identify the highest-value valorization routes.
Potential applications include fractionated functional fibre and polysaccharide products, bio-based material or composite feedstocks, and engineered adsorbents or biochar. The workflow explicitly accounts for the stream's variable composition, low pH and residual nitrate, and ranks candidate routes by value, feasibility and robustness to variability before committing to physical prototyping.
Technical specifications
How it works:
- Feedstock characterization: composition, variability, pH, nitrate and moisture are quantified to establish the design space and agree value and scalability criteria.
- Predictive modeling: domain-guided AI models link feedstock characteristics and process conditions to candidate product properties and yield.
- Active-learning design of experiments (DoE): high-value routes are targeted with minimal pre-treatment, reducing experimental burden and accelerating route selection.
- Lab- and pilot-scale prototyping: the top 2-3 routes are prototyped and characterized across representative batches, measuring product properties, yields and processability.
- Down-selection and scale case: results are fed back into the models to converge on 1-2 lead routes with a defined scale-up path, validated samples, a data pack and a reusable model.
The platform uses a centralized data foundation, domain-specific AI models, and explainable insights to replace manual, spreadsheet-based R&D workflows. It includes over 35 domain-guided models trained on polymers, chemicals and advanced materials, enabling R&D teams to predict material properties and optimize formulations with greater speed and confidence.
Technology readiness level
The underlying platform is commercially deployed and has been validated in adjacent materials programmes. In a paints programme it modelled more than 50 ingredients against 15+ properties and cut model error in half, from 16% to 8%; in another programme it delivered a production-ready phase-change material more than 65% faster than traditional R&D.
For the pectin side-stream application, the workflow is at early-stage validation. The proposed programme covers scoping, route screening and modelling, lab prototyping, and down-selection to build the value and scalability case, with subsequent phases planned to advance readiness through pilot-scale characterization and validation.
About Polymerize
Polymerize is an AI-native platform specifically engineered to transform materials research and development. It operates as a system of intelligence that integrates a unified data foundation, domain-specific AI models, and explainable insights to accelerate the discovery, development, and scaling of new materials. By digitizing experimental workflows, the platform replaces manual, spreadsheet-based processes with a centralized, data-driven backbone. The solution features over 35 domain-guided models trained on polymers, chemicals, and advanced materials, which allow R&D teams to predict material properties, optimize formulations, and gain scientific explanations for AI-driven recommendations.
The platform serves global R&D teams across industries such as automotive, specialty chemicals, coatings, and packaging, helping them reduce failed experiments and shorten go-to-market timelines. By capturing and leveraging institutional knowledge, Polymerize enables organizations to turn years of trial-and-error into targeted innovation, ensuring that data from past experiments informs future research. Customers use the platform to achieve significant ROI through increased discovery speed and streamlined operational forecasting, with security compliance standards like SOC 2 and ISO 27001 integrated to protect sensitive intellectual property. Founded in 2020 and headquartered in Singapore, the company maintains a global presence with offices and strategic collaborations spanning Asia, Europe, and the United States.
delivered to your inbox.