Get expert insights from Chris Farrow, the VP of Materials Science Solutions at Enthought, a leading provider of scientific computing solutions!
Are you interested in unlocking the full potential of your materials data? Whether you’re developing new materials for cutting-edge applications or optimizing existing materials for better performance, you need advanced tools and expertise to manage and analyze your data effectively. This is where Enthought comes in. Enthought is a leading provider of scientific computing solutions that help organizations across industries to accelerate innovation, reduce costs, and increase productivity.
To learn more about Enthought’s materials science solutions and the challenges and opportunities of materials data management and analysis, we interviewed Chris Farrow, the VP of Materials Science Solutions at Enthought. Chris has more than 20 years of experience in materials science, chemistry, and software development, and has been instrumental in creating and advancing Enthought’s materials science platform. In this interview, Chris shares his insights on the latest trends in materials science, the role of machine learning and AI in materials data analytics and the benefits of a unified data platform for materials innovation.
Let’s dive into the interview and see how Enthought can help you solve your materials data challenges and achieve your innovation goals!
Kindly brief us about yourself and your role as the Vice President, Materials Science Solutions at Enthought.
I am a lover of science and technology. I spent some time studying physics and mathematics in school, not quite sure what I wanted to do, before I eventually decided to pursue a PhD in physics at Michigan State University. That’s where I started using Python as part of a grant that was focused on creating data analysis tools for studying materials. It was that experience that really cemented my passion to make an impact by creating software.
I’ve been in industry for about 16 years, and started as a developer at Enthought in 2011. I transitioned to serving as a consultant and now currently lead Enthought’s Materials Science Solutions Group. As VP of the group, I oversee digital transformation solutions for specialty chemicals and semiconductor industries, as well as the development of novel technologies for materials data management and discovery.
Please share your source of inspiration for exploring various facets of technology.
I learned early in my academic career that I could get more done at the computer, and importantly, there were some things that could only be done by a computer. As I progressed through graduate school, I started writing applications for other scientists to use, which really intensified my desire to make an impact with my work.
Once scientists start using your software, you become their go-to resource in that science, which is very rewarding. This led me along a path where learning and creating became intertwined. That drive to learn and create is my inspiration. I think about advances in battery materials, for example, and I ask “What is driving improvements to storage capacity?,” “How do they measure that?,” “Can it be automated?,” “What if we could predict X?” Scientific curiosity inevitably leads to technology questions.
Please brief our audience about Enthought and give us an overview of its standout solutions.
Enthought is a globally recognized leader in scientific computing, providing specialized solutions that accelerate scientific innovation across various industries. We partner with science-driven companies in the electronic, semiconductor, materials design, manufacturing, pharmaceutical, biotechnology, energy, and consumer goods industries.
Our transformative solutions, from AI-assisted interpretations of subsurface seismic data to quantum simulations for material informatics and ML models for cancer therapeutics, have helped businesses achieve breakthrough discoveries in record time. Scientists all over the world also use Edge, our cloud-native platform that serves as a central hub for all their R&D data, analysis, and application needs. We also have programs, like our Materials Informatics Acceleration Program, that upskill scientists with the new digital skills needed to leverage technology to make new discoveries.
What are the core values on which Enthought is formed and what is the mission of the organization?
Enthought’s mission is to help companies fully realize their business objectives and gain competitive advantages by digitally transforming their R&D organizations, from idea generation to custom software to empowering teams. Ultimately we aim to help companies answer the question, “What could be accomplished if your scientists could spend 100% of their time advancing their discoveries?” We have a deep understanding of the complexities of science-driven processes and scientific data as well as advanced computing techniques.
We also bring a unique data-centric approach that encourages R&D leaders to think differently about how they conduct science and expand what’s possible in the lab. This approach and understanding allows us to conceptualize and deliver solutions in a way other firms cannot.
Being a thought leader, how do you strategize to bring to light Enthought’s mission and vision?
First and foremost, we want to make an impact with our work. Given the historical hype around AI and ML, this is critical. You can only go so long building models and demonstrations before someone asks what it’s good for. Additionally, we’re driven to empower scientists. 90% of our global technical team have advanced STEM degrees, with 65% holding Ph.D.’s. So we understand scientists’ challenges, their goals, and the pressures they face. Our solutions are developed by scientists for scientists, while focused on what brings value to the enterprise.
As far as markets, we partner with companies of all sizes and stages, from Fortune 500 to startups, in science and innovation-driven industries like materials science, pharma, and chemistry. We build custom solutions built around their niche so they can do more, accomplish bigger things. I mentioned batteries earlier. The battery industry is highly scientific, experiencing tremendous development and commercial activity, and needs digitalization. We also help companies leverage new technologies like material informatics (MI), which is poised to change how materials and chemicals R&D is performed.
Can you explain to our audience how MI helps powerhouses like Resonac make better, more efficient decisions in the lab?
Global materials and chemical companies have faced increasing pressure to continuously get new products to market faster, and material informatics, a data-centric approach to innovation, has emerged as a strong solution to help companies establish and maintain their market presence. It can be difficult for an organization to keep up with the cutting edge trends in MI where new approaches are constantly in development. This is especially true of smaller companies, and companies in Japan, where it can take years for a new technology or technique to become widely accessible.
Chemicals supplier Resonac (formerly Showa Denko) has always been at the forefront of change and understands the potential of MI. To maximize and accelerate their efforts, Resonac partnered with Enthought to uplevel their capabilities and bring their computational solutions into product development across the company. Within six months, we’ve upskilled their scientists in the latest MI techniques, while co-developing solutions for real challenges they’re facing. These solutions are giving them an advantage in capability and efficiency that puts them years ahead of their competitors. They are now on a new trajectory of capability, at multiple levels, creating exponential value.
What is your perspective on the significance of enhancing AI and ML in such a manner that they can be fully utilized, rather than being a replacement for human abilities?
It’s really hard to replace a scientist’s skill and intuition for their science. AI and ML are not going to replace scientists wholesale. However, there are abundant use cases demonstrating how AI and ML can remove mundane work, and enable them to make better use of their highly specialized skills. When AI and ML are used effectively, they don’t change individuals’ work, but rather they change their relationship with their work.
We’re starting to see the same pattern emerge in multiple places in our MI Acceleration Program. For example, we’ve helped scientists build an experiment recommendation tool, essentially a machine learning model and Bayesian inference system to guide them while developing chemicals or materials with target properties. These systems can make use of historical data in a statistically rational way that humans cannot. The tool guides the scientists to an optimal formulation or mixture much faster than intuition alone.
Recently, a client told us about how they tested the solution in the lab, and the scientist didn’t follow the recommendation. Instead, they used the recommendation as a starting point, and reasoned about how to improve the recommendation. When they followed their ML-inspired recommendation, they hit their target on the first try. This is exactly how it should work. AI and ML helps us see what we cannot, which helps us make better solutions.
How important is it to leverage the power of AI in order to boost business performance?
It’s very important, but it matters how you do it. Organizations often fall into the common trap of thinking that the integration of AI or ML on top of their current data management processes and workflows will bring immediate value to the business. Research has shown, however, that the majority of initiatives fail when they take this technology-first approach. While it’s clear that companies today need to embrace new technologies to be competitive, to leverage the true power of AI, they also need to invest in the people and processes around them.
When done effectively, a business should fundamentally change. Take the experiment recommendation system I just mentioned. We’re seeing the time it takes to formulate and deliver a material to spec being cut in half, or more. People are using AI and ML to generate and screen molecules with specific properties. Natural language processing is being applied to literature search and for discovering whether a material has been patented, simulation, a staple of materials science.
Work is accelerated by surrogate modeling. You can buy or develop software with this capability, or enable a scientist with digital skills to build it with some help. It won’t be too long before companies who are leveraging AI and ML are outcompeting those who do not.
Could you provide our audience with an overview of the latest developments in the new generation and your strategy to meet the evolving requirements of the AI/ML infrastructure?
We’re already seeing a few trends we think will take center stage for scientists this year. As you know, ChatGPT has of course made waves across industries already. For science in particular, at Enthought, we’ve long believed the role of tech should be to empower scientists and accelerate their work. Chatbots like ChatGPT that understand free-form or structured input, and can turn it into detailed information, is a perfect example of the sort of tool that can help free highly skilled scientists from low-value work.
At Enthought, we specialize in helping companies employ the solutions that will cut their research times in half. We’ll be keeping an eye on ChatGPT to see how it evolves and could potentially play into science.
Another trend we’re seeing is in the shift to data-centric deep learning. The existing dominant paradigm is model-centric – you fix the data you’re working with and tweak your model to get a good fit. The data-centric approach is flipped, with the focus on data quality once you have a reasonable model. This approach aligns more with how we approach scientific data in general. After all, if your measurements and analysis are inconsistent, the rest of the features aren’t going to be useful for machine learning or much of anything else.
Whether it’s AI, laboratory transformation, or something else, organizations must ask themselves, “What is the quality of my data and how do I go about improving it?” We can help companies answer that question so they’re implementing purposeful AI.
We’re also seeing new architectures mature, like graph neural networks, that will have a big impact on materials research. Success cases for this are still emerging, but I’m excited by what I’m seeing.
Enthought is well equipped for fulfilling the dynamic needs of the AI-ML infrastructure. Our MI solutions have a baked-in virtuous cycle that keeps us ever-ready to take on the next challenge. We hire from leading materials science and MI programs throughout the world, and through our MI Acceleration Program, these experts are continuously exposed to the real-time data needs of our clients.
What advice of value would you offer to emerging entrepreneurs and professionals in the industry?
Prior to joining Enthought, I was a scientist with no business sense. Entrepreneurship came to me last out of all of the things I learned. When my management responsibilities started leaning more into growing the business instead of just running it, I started reading. It’s not hard to find top-ten reading lists for would-be entrepreneurs. Read about your industry as well to learn who the big players are, what made them successful, what their failures were. Listen to podcasts to get different perspectives and get inspired. Even though we’re all busy, it’s definitely worth investing your time into learning new things.
How do you plan to scale up Enthought’s growth curve in 2023 and beyond?
Despite the economic landscape, companies are continuing to invest heavily in digital transformation to gain a competitive advantage. We recently announced a new 5-year extension of our partnership with one of the largest semiconductor manufacturers in the world, Tokyo Electron, as well as new partnerships with specialty materials company TBM and chemicals supplier Resonac.
As innovators in their fields, they know building a strong technology infrastructure and data-centric culture now will set them up for decades of success in the future. We’re also seeing that R&D organizations are looking more to purpose-built solutions after experiencing failed initiatives that started with either a software solution that over-promised and under-delivered or a technology-only approach. In fact, we’re having a lot of conversations around that right now. Now more than ever, R&D organizations are recognizing the importance of holistic approaches and solutions.
Vice President of Materials Science Solutions at Enthought
Chris Farrow holds a Ph.D. in physics from Michigan State University and degrees in physics and mathematics from the University of Nebraska. Chris has spent 16 years working as a physicist in materials discovery and characterization. Chris currently leads Enthought’s Materials Science Solutions Group, where he oversees Digital Transformation solutions for the Specialty Chemicals and Semiconductor industries, as well as the development of novel technologies for materials data management and discovery. Chris is based at Enthought’s headquarters, in Austin, Texas.