I currently serve as a Senior Director on the NVIDIA government affairs team, where I lead a group that advises on all technical issues. In this role and all throughout my career journey I have operated at the intersection of technological innovation, public policy, and geopolitical strategy. I have often advised on legislative, regulatory, scientific, and industry outcomes for a variety of different technical topics including advanced manufacturing, supply chain resilience, robotics, quantum technologies, biomedical devices, semiconductors, and more recently artificial intelligence.

There are always misunderstandings and miscommunications between the scientific, commercial, and government sectors based on the different incentive structures, priorities, and relevant timelines associated with those different communities. The challenge is that the best set of decisions requires insight from each of these perspectives. My mission has been to learn the language of each community, understand the different factors that inform strategic decision making, and thereby offer a harmonized set of recommendations for the best national innovation strategy.

This is the approach that I brought to my previous roles such as my efforts with defense semiconductor supply chain modernization at DARPA or my contributions to the authoring and passage of the 2022 CHIPS and Science Act at the Semiconductor Industry Association. It remains my guiding philosophy as I help navigate these same relationships from within my role at NVIDIA in pursuit of the best national strategy for the artificial intelligence revolution.

I am a scientist by training, specializing in the materials that make modern computational technology possible – semiconductors and related materials. I came to Washington, DC for a post-doctoral fellowship at the Navy Research Laboratory but despite my interest and passions at the lab bench, I found that there was an enormous need for science communicators to help bridge the gap between the scientific and policy communities. From there, I pursued a policy fellowship in the White House Office of Science and Technology Policy assisting with the National Nanotechology Initiative, with a particular focus in the future of computing technologies. After wrapping up that fellowship, I joined DARPA to assist with a new set of programs focusing on defense microelectronics and semiconductors called “The Electronics Resurgence Initiative”, which sought to craft a closer relationship between the defense semiconductor community and the innovations that were emerging from the commercial and industrial sector. This was in 2018, but there was a growing urgency to the geopolitics of semiconductor manufacturing and in particular the ways in which the United States had fallen somewhat behind in certain stages of the semiconductor supply chain. This all came to a head during the COVID supply chain disruptions of 2020 and 2021, which helped motivate the creation of what came to be called “The 2022 CHIPS and Science Act”. By the time I had reached 4 years at DARPA, this legislation was starting to draw serious interest on capitol hill. I felt that my experience in navigating scientific, policy, and industrial stakeholder groups would add value to this legislative process and so I joined the Semiconductor Industry Association just as the bill writing process was getting underway. By almost pure coincidence, it was around this time that advances in artificial intelligence were really starting to catch the attention of DC, which motivated my move into the government affairs office of NVIDIA – this was an opportunity to synthesize my prior experience across multiple roles to help solve the defining challenge of this generation: to successfully navigate the AI technology revolution from a scientific, economic, and policy perspective.

It began purely as a fascination with the physics of computing technology that I felt as far back as middle or high school. The idea that you could make something that seemed to function as if by magic with a relatively mundane set of materials and electrical principles really captured my imagination. My experiences building my own gaming computers in the late 1990s and early 2000s certainly helped sustain this fascination. And I did in fact use NVIDIA chips. As I explored this topic through my university education in physics and materials science, there was a sentiment even 20 years ago that the dominating principle behind computational progress, Moore’s Law, was nearing the end of its useful lifetime. And this was very exciting to me from an innovation perspective – it meant that the aperture was soon to open on new designs and new material possibilities within the compute stack. This ultimately progressed more slowly than I might have guessed as an undergraduate – Moore’s Law remained a dominating factor far longer than some expected. Even so, as systems continued to improve, as the smart phone revolution happened, it felt to me through the 2010s that we were on the precipice of some enormous technological change which would be driven by the pace of innovation in the computing sector. And it seemed that this change would drive enormous increases in scientific discovery, in economic output, and in quality of life. I wouldn’t have described it as “artificial intelligence” necessarily at that point, but it became clear in 2022 that AI was that looming wave of innovation I had felt coming in those prior years.

In the near term, AI is going to impact innovation in the economic sectors that get the most use out of “intelligence”. It’s obvious that any software-defined business fits within that definition and I think it’s pretty unambiguous that the ability of AI to create enormous amounts of high quality software will be the first major economic and innovation impact. Beyond that, I think there are several sectors which will benefit first from integration of AI into less abstract applications than software. Embedded intelligence in the telecommunications stack seems like an enormous source of innovation within the next 5 years. Autonomous vehicles, manufacturing, and advanced robotics, or what we often call “physical AI” are likely to see innovation boosts in that timeframe as well. I am very optimistic about the integration of AI into the scientific enterprise, but 5 years is a short timeframe to expect major changes. The impact of AI on scientific discovery will be enormous, but we should be realistic about timelines.

Government has an important role to play in advancing responsible AI or any technology use – simply their convening authority is important enough as a stand alone role. Setting standards and creating unambiguous rules of the road for technology deployment is a key part of the government role as well. Finally, one of the major roles that government has to play on advancing responsible use of any technology during a major technological revolution is to create programmatic support that ensures small and medium players don’t fall behind the curve. In all cases though, industry is the implementer of innovative technological solutions. That means that while the government’s role is to create standards or identify desired outcomes, the industry role is to decide the best way to get to those outcomes. There are risks of unfair rules or even regulatory capture when the government is over prescriptive on how to solve specific technical challenges.

The biggest challenge that is rooted in both technology and in policy is the geopolitical reality of modern technology supply chains and the overly optimistic expectations that some nations have about decoupling from international technology dependencies. The reality is that modern supply chains, especially for the most advanced technologies, are highly complex and interdependent, crossing multiple national borders many times before a product makes its way into the hands of a consumer or an into an enterprise. This is not something that a simple set of policies can fix, nor is it even something that is necessarily desirable to change at scale – the miracle of innovative modern technologies and the miracle of relatively open international technology markets are linked to one another.

It’s important to understand the limitations of those who help build out the connective tissue and communication networks between stakeholders and communities – my role is to make sure that the final decision maker in government or in industry has the most complete and correct set of information at their disposal when it comes time to make a decision. My work helps strengthen innovation and technology leadership because I make sure that government decision makers understand the things that only industry can really know – how is technology infrastructure deployed, how do we build innovation ecosystems, what are the supply chain limitations that we encounter, what market competition do we face around the world? If a set of policy decisions are made, what is likely to be the industry implementation or reaction? This works in the other direction as well. I help decision makers at NVIDIA understand what information the government sees – how are their voters and other stakeholders thinking about AI, how does the national security community look at technology risks, what geopolitical competition do we face around the world? In order to create the strongest possible innovation and technology leadership we need a mutually predictable and respectful relationship between industry and government.

There’s no simple set-and-forget answer to this question unfortunately. Starting with good data management, strong security practices, and an emphasis on privacy will be critical for almost any application. AI models are not like conventional software in that they should be performing better over time rather than offering a static capability – this means that continuous updating and validation will be required with a virtuous cycle between model use and data creation. For the most critical applications, AI may be most appropriate as an augmentation of human decision making rather than a replacement.

I am extremely excited about closer integration of optics into the core of the computing stack, the so-called “co packaged optics” have an enormous potential to increase the scale up capabilities of AI infrastructure on top of the enormous scale up that has already been successfully achieved. It’s a very exciting time in the optical world and systems that we might have guessed wouldn’t see commercial deployment until well into the 2030s are now seeing integration onto product roadmaps by 2028 or 2029.

My leadership principles are rooted in my background in science and in the use of data and information to arrive at the best set of decisions. Because the topics my team engages on must combine extremely diverse information sets from technology roadmaps, investment statements, economic projections, and even sometimes geospatial imaging, it is very important that we give great diligence to crafting our decision-making process and ensuring we have the best available data. For myself, that means that my key principles are transparency and responsibility with my process and an openness to making changes when new information becomes available. For my team, that means maintaining technical competence, having a freedom to operate flexibly within domains of strength, and a confidence in the quality of our conclusions that comes from strong underlying process and the best available data.

No matter your future career path – the most successful young professionals today will be the ones who use AI tools to make themselves smarter and more effective. These tools do not replace true human expertise, rather they help round out the skill set for those skills outside of your expertise. We are entering an era where domain expertise in the hands of somebody proficient with AI tools will become more valuable than ever before. As a student just beginning your journey, this is the perfect time to follow your interests and curiosities, turn passion into expertise, and find success through the integration of that expertise with these emerging AI tools.