Featuring Associate Professor Aaron Stebner

1. What is your field of expertise and at what point in your life did you first become interested in this area?

I work at the intersection of mechanics, metallurgy, machine learning, and manufacturing. I became interested in engineering as a small child – my grandfather was an engineer, and when I would spend time with my grandparents in the summer, I would go to work with him, and I was fascinated with drawing boards, alligator clips, circuits, and more. In high school, I started interning at the business he had built that primarily developed automation and test equipment for circuit breaker manufacturing (he had passed and my uncle then ran it). I started in the stock room, worked through the machine shop, assembly, and into quality control in my first years there. Then I became an engineering assistant as I went into my undergraduate studies. I had thought that I wanted to be electrical engineer (like my grandfather), but after 6 – 8 months of assisting EE, I realized that my true passion was in mechanical engineering, and I moved over to ME – so that foundation instilled in me that I had a passion for ME, manufacturing, and automation. The metallurgy came years later, when I won a graduate fellowship to work at NASA Glenn while earning my Master’s degree. I worked with metallurgists there who were developing new shape memory alloys, which fascinated me. I resisted materials science and metallurgy for many years, insisting that should be someone else’s job, and I should stick to manufacturing and ME. However, it became evident that you can’t engineer with shape memory alloys or develop their manufacturing unless you deeply understood their metallurgy – that resonated with me when I attended a conference in 2008 while working for a startup company that was commercializing some of the new shape memory alloys the group I’d worked with at NASA had developed. When I returned from that conference, I signed up for my PhD program the next week and dove deep into the intersection of metallurgy, manufacturing, and mechanics. The machine learning came years later, several years into my faculty career. We were working with several companies and the state Office of Economic Development in Colorado (I started my faculty career at Colorado School of Mines) to develop an R&D center and technology incubator to support the growing metals 3D printing industry. When I asked the industry people why they needed a center/consortium at Mines in this area – what were they not getting at other additive manufacturing centers at that time (this was 2014/2015), they said “no one is helping us with our data problems.” So, that became our mission – data informatics innovations in metals additive manufacturing. Here at GT, I’m thrilled by the opportunities, colleagues, and infrastructure available to bring it all together – our big vision for this IMat initiative is to develop R&D test beds and technology incubators for AI materials manufacturing.

2. Why is your theme area important to the development of Georgia Tech’s Materials research strategy?

Largely, our materials research laboratories (nation-wide and globally, not just at Georgia Tech) have been designed and built to support human operators. However, AI cannot independently function in the same way and in the same environments – or, at least, we will never realize its full potential if we make it play by our rules. Re-thinking and designing new materials laboratories that can operate autonomously and semi-autonomously is critical to be at the forefront of future innovations.

3. What are the broader global and social benefits of the research you and your team conduct?

Lowering barriers and times for the discovery and development of new materials and manufacturing – lower costs, faster times to deployment, increased sustainability, and finding better solutions. Also, with AI engines, the ability to distribute manufacturing to local/underserved parts of the globe and our nation – we saw this at the onset of COVID – when our corporate supply chain was unprepared to meet the demand, people were able to contribute respirators, masks, and more using the 3D printers in their garages, libraries, schools, universities, and hospitals and serve their community. However, people in their garages are rarely equipped to qualify/certify/ensure safety of critical parts and widgets on their own – the data infrastructure + AI enables qualification/certification to happen through statistics, and then rapid dissemination of the manufacturing “how to”. One could even imagine a future where the burden of qualification and certification could be shared across everyone participating in the supply chain – that will take a lot of policy and economic reform and rethinking as well, but as we gain confidence in our understanding of statistical models and data management infrastructure and software, it becomes more and more feasible.

4. What are your plans on engaging a wider GT faculty pool with IMat research?

I think the group of involved faculty now spans 7 or 8 schools and 3 colleges, at least – I’ve stopped counting, to be honest – the interest and support of colleagues here at GT is tremendous. On our larger proposals, there are anywhere from 20 – 30 faculty involved – I think this next one we may exceed 40. I welcome anyone who has ideas for how they can contribute or wants to learn more about the vision for AI materials + manufacturing test beds to email me anytime, and we’ll setup a time to meet and discuss. I also intend to hold some workshops and conferences – we received funding to start a consortium that will hold quarterly meetings for any interested business or faculty, and newsletters will also be sent, starting in 2022.