Ry Nduma is a second-year undergraduate in the Department of Materials. Ry is completing a placement with Professor Aron Walsh as part of the Undergraduate Research Opportunity Programme. During the placement, he is exploring how to develop smarter, faster ways to discover new materials using AI and programming.
Can you tell us more about your placement?
When I first started my UROP placement in the Department of Materials, I wasn’t sure what to expect. I had found myself at a crossroads and needed clarity on what I hoped would materialise out of my degree and time here at Imperial. So, I decided to go back to the roots of which I was interested in: programming. With the recent developments in AI, I was curious to learn how to bridge AI with my degree in Materials Science and Engineering.
I emailed Professor Aron Walsh, introducing myself and sharing my programming experiences from our first and second-year Materials modules, alongside what I had learned from online courses. To my delight, Professor Walsh welcomed me into his research group for the placement, with a project focused on computational materials design.
What did your project involve?
My project was part of a larger effort to develop smarter, faster ways to discover new materials. Before starting the project, I had to brush up on the basics and become familiar with the group’s existing research. First, I looked at the fundamentals of materials informatics, understanding the simple rules chemists use to predict how atoms behave and learning about crystal structure databases.
My first task involved working with SMACT (Semiconducting Materials by Analogy and Chemical Theory), where I developed simple rules for screening inter metallics based on chemical pattens and behaviours. This hands-on introduction taught me how decades of chemical knowledge could be encoded into algorithms that screen millions of possible materials in minutes, something that would take humans years to do.
As I grew more comfortable using computational tools, I enjoyed learning more about how artificial intelligence, particularly large language models (LLMs), are beginning to transform materials discovery. Traditionally, designing new materials involves extensive and expensive trial-and-error experiments like synthesising compounds, testing properties, and iterating based on results. With AI becoming more powerful, it’s exciting to think about using all the chemical knowledge built into these models to predict how materials will behave—and even help design them—before we step into the lab.
After this, I worked closely with my supervisor and started developing some experiments probing into how large language models (LLMs) have learned to understand the complex relationships, rules, physics, chemistry, and mathematics that underpin materials design. We discovered that because these models have been trained on huge amounts of scientific papers and text, they are able to learn and identify vast patterns and knowledge that took humanity centuries to discover. I began investigating how this accumulated knowledge could be harnessed to design, synthesise, and discover new materials with tailored properties, from next-generation solar cell absorbers to novel battery electrodes. The potential applications seem limitless: materials that could capture carbon from the atmosphere, superconductors that work at higher temperatures, or catalysts that make chemical processes more sustainable and efficient.
Why did you want to complete a placement in this research field?
Science and computers have always been fascinating, but materials informatics offers something unique: a chance to apply computational thinking to tangible, real-world challenges. I found it exciting that this field combines programming and data analysis with materials science to tackle practical challenges, from designing better batteries, catalysts, and semiconductors for energy and electronics, to advancing sustainable manufacturing.
During my placement, I’ve been building confidence in using computations tools designed specifically for materials science, like SMACT for high-throughput screening, neural networks like MACE that can learn interactions between atoms in a system and accelerate atomic-scale simulations of material properties for virtually any kind of material, and newer generative AI tools like Chemeleon that open up exciting possibilities for designing new materials.
But beyond the technical skills, I am continuing to gain experience in the research process and how to think like a researcher, whether it’s formulating hypotheses, designing computational experiments, and interpreting results considering what we already know about materials. One of my biggest lessons has been learning to navigate the delicate balance between computational predictions and experimental reality, understanding that our models are only as good as the physics we encode in them.
What did you enjoy and how will this help you in future?
This experience really changed how I see my future. Before the UROP, I thought of programming and materials science as separate interests, but now I see them as complementary tools for tackling humanity’s biggest challenges. The confidence I’ve gained, not just in technical skills but in my ability to contribute to cutting-edge research, has made me certain that I want to pursue a PhD.
I also attend the group’s weekly group meetings, where the Walsh group encourages a culture of asking questions, sharing knowledge freely, and having open conversations. I’ve realised that research isn’t a solo effort; it’s built on teamwork, discussion and sharing ideas. This experience has been transformative for me.
The UROP placement has been more than just a research experience for me: I have learned how combining ideas from different fields, like chemistry, programming, and materials science, can lead to faster and smarter ways of solving scientific problems. What excited me the most was seeing how AI and materials science can make real contributions to sustainability, energy, and even broader human progress
What advice would you give to students who want to apply for a placement next summer?
When contacting professors, be specific about why their research excites you and how your background, even if your experience is limited, could contribute to their research group – you’d be surprised how far a little self-belief can take you. Also, embrace the steep learning curve. You may be overwhelmed in your first weeks, surrounded by concepts and techniques you’ve never encountered. This is normal and thankfully it doesn’t last forever! Be patient with yourself, ask questions, and remember that every expert started in the same place. Finally, keep a growth mindset. The skills you lack today can be learnt tomorrow, and the questions you can’t answer yet might lead to your biggest breakthrough. Good luck!