Bridging R&D and Data Science – A Conversation with Maxwell Dylla

🔑 Key Takeaways

1. Applying Machine Learning to Battery Research

Maxwell’s work at Cuberg highlighted the power of machine learning in accelerating battery formulation. By using active learning loops and response surfaces, his team was able to efficiently test and optimize electrolyte compositions. The ability to predict optimal formulations using data-driven models significantly reduced the trial-and-error process, leading to faster breakthroughs in high-performance battery technologies.

2. Building a Traceability Solution for R&D and Manufacturing

One of the major challenges in scientific R&D is managing data across research, production, and supply chain processes. Maxwell tackled this problem by implementing a robust traceability system that integrated data from Cuberg’s R&D experiments, pilot manufacturing line, and supply chain. His team leveraged tools like Trino for cross-catalog queries, Prefect for ETL orchestration, and a lakehouse architecture to streamline data access and analysis. This approach ensured seamless data flow between teams and enabled data-driven decision-making throughout the battery development process.

3. Open Source Possibilities in Battery Analytics

Maxwell recently introduced "Battery Pulse," an open-source project aimed at harmonizing test data from battery testers. The project provides a standardized way to collect, integrate, and analyze test data, addressing a critical gap in the battery industry. By making this tool available to a broader community, Maxwell is unlocking speed to deliver and efficiency gains to researchers, startups, and manufacturers looking to improve their data workflows.

💡 Reflections

The importance of combining an analytical mindset, domain knowledge with data skills:

- Maxwell emphasized how his academic background in material science and physics provided him with valuable causal thinking and experimental design skills that applied well to data science.

- He also noted the need to understand both the technology and the business context when working in data-driven industries like battery development.

Data engineering plays a critical role in enabling scientific research and R&D-driven companies to scale their operations:

- Maxwell highlighted the challenge of integrating disparate data sources from R&D, manufacturing, and supply chain at Cuberg.

- He discussed how they leveraged tools like Trino, Prefect, and a lakehouse architecture to create a centralized, queryable multi catalog data repository to support data-driven decision making.

Lack of pre-built tools - leading to opportunities for standardized solutions

There is  potential for open-source tools to standardize and accelerate data integration in specialized domains:

- Maxwell's "Battery Pulse" project aims to provide a standardized solution for integrating battery testing data.

Battery Testing Data

‍

Data engineering plays a critical role in enabling scientific research and R&D-driven companies to scale their operations. The use of scalable data transformation workflows, structured data storage in cloud platforms, and open-source solutions can help research teams overcome bottlenecks and focus on discovery rather than data management. Companies adopting these best practices can enhance collaboration, improve regulatory compliance, and accelerate time to insight.  Maxwell’s “Battery Pulse” project will provide significant value to players in this vertical.

Stay tuned for more stories from inspiring data leaders in our series.

‍