Onward, with an extra helping of secret sauce.

As they say, about all good things…my time has come to an end at Microsoft. After 12 years spent accelerating impact through strategic research collaborations in Microsoft Research, the time has come to move on. So, if you’re reading this, I hope you’ll indulge me on my personal and professional stroll down Microsoft Memory Lane. For those of you looking for the TL;DR (Too Long Do Not Read Version): It was a great ride and I’m especially grateful for the opportunity to meet and work with so many kind and brilliant people.

My Microsoft journey began when the magnanimous Lee Dirks took a risk on me. He brought me in as an “outsider” to lead a small team of outreach experts with a focus on education and scholarly communication related projects. Lee tragically passed away just six weeks after I started but anyone who was lucky enough to know him, even for a short time like me, knows what a great force of nature he was. While I was coming from another leadership role where I managed a team of 12 grant managers, I was not prepared for all the high achievers on my new team who had all been with the company for 10 – 12 years already. Immediately I felt the imposter syndrome kick in. Luckily, they were all (mostly) patient and helped me with my own learning curve. Shout out to Donald Brinkman and Rane Johnson on that front. On my first day I remember Tony Hey, our team VP at the time saying to me, “Ok, Roy, now you’re going to tell me what Microsoft Research should be doing in the education space.” Talk about pressure.

The first project I led for Microsoft was ChronoZoom. I still think this technology was ahead of its time. It was the brainchild of the amazing Walter Alvarez and Curtis Wong and never would have happened without Roland Saekow’s vision and unfailing belief in the project. I didn’t realize it at the time, but this was my first taste of the power and potential of MSR + academic collaboration to turn ideas into action.

My next big assignment was to lead the team driving the development of Microsoft Academic. This was an engineering effort meant to generate an early version of a scientific research digital assistant and while the ideas were bold and the people involved were exceptionally talented, the project was discontinued. I am particularly grateful to Alex Wade who was an exceptionally supportive and collaborative partner in this endeavor.

Another big swing was Project Tokyo, which focused on pushing the boundaries of AI to create technologies that could support the needs of blind athletes competing in the Tokyo Paralympic Games. This project has since evolved into an exciting line of ongoing research for AI-driven visual aid technologies. Cecily Morrison and Ed Cutrell are perhaps two of the kindest, smartest people I met in my time at Microsoft and this project allowed me to work closely with them both. A real gift.

Then there was the Urban Innovation team where I got to know Scott Counts, Asta Roseway and others as we developed the first of what would be many workshops I led meant to foster internal and external collaboration. This workshop was looking to better understand disparities in urban settings. Madeleine Daepp was a graduate student at the time, attended the workshop, and later told me she joined Microsoft Research at least in part thanks to that gathering. She’s gone on to do some incredibly important research recently on election related misinformation.  

Another exciting part of my journey was the opportunity to work globally. The Homomorphic Encryption Standardization Workshop catalyzed a new community of like-minded experts. This time we pulled people together from Microsoft and academia as well as Samsung, Intel and several start-ups to successfully drive the creation and adoption of a new international standard for homomorphic encryption. Again, I was given the gift of new, smart people to meet and work with inside the company like Kim Laine, outside the company, like Kurt Rohloff and those who formerly worked at Microsoft, like Kristin Lauter. (Gold star for anyone who can find me in the photo below of our first workshop attendees.)

The Cortana Intelligence Institute was a partnership of a new kind of academic collaboration with RMIT University in Melbourne, Australia working closely with many RMIT faculty and students.  Ryen White, who helped conceive and manage this program with me, has been both mentor and friend and led the technical side of this program brilliantly with heavy lifts also from Ahmed Awadallah and Paul Bennett. This project drove the creation of new AI capabilities for digital assistants directly informing Microsoft products.

The latter half of my time in Microsoft Research allowed me to focus on social impact, an area I am particularly passionate about. This really came into focus when the pandemic hit, and the entire company seemed to shift into a massive response mode. Microsoft is lucky to have Eric Horvitz as the Chief Scientific Officer, and I was lucky enough to partner with Eric along with Chris White and my colleague Kenji Takeda on the Studies in Pandemic Preparedness program that sought to quickly establish new academic partnerships to dramatically accelerate the pace of discovery of new science related to the pandemic.

As the Generative AI tsunami started to crest, I found myself working more and more on issues at the intersection of large language models and social impact. The Microsoft Turing Academic Program was an early effort at harnessing this new potential for good and evolved into the Accelerating Foundation Models Research program. It was a truly awesome experience conceiving of and co-leading the development of this early work with Dean Carignan, one of the best collaborators I had the pleasure to work with at Microsoft.

Microsoft also made bold proclamations in response to the climate crisis. The Microsoft Climate Research Initiative is still helping to accelerate priority research for Microsoft and customers to achieve sustainability related goals faster. Karin Strauss could not have been a better partner in conceptualizing and implementing this important program. She is one of the kindest and most technically proficient researchers I have ever met. Bichlien Nguyen and Jake Smith, key Microsoft researchers in this initiative, also made managing complex technical partnerships look easy.

With more and more of the company focused on the potential global impact of generative AI, I once again found myself turning to a new domain with the opportunity to engage with a whole new cast of researchers. I was asked to look for ways to examine how Generative AI was going to affect the way people think and work. Thus, was born AI, Cognition and the Economy, or AICE. This time around, I had the good fortune to work more closely with Abi Sellen who first crystallized for MSR the need to study that arc of how Generative AI is going to change the way people think and what that means for workers, jobs and ultimately global economies. I also feel fortunate to have had the chance to get to know and work with some of Microsoft’s top social scientists including Dan Goldstein, David Rothschild, Jake Hofman, Sonia Jaffe and Eleanor Dillon on a series of AICE workshops bringing together top minds from around the world. Ultimately, AICE led to a round of funding for exploratory pilot collaborations that I managed to better understand things like how AI can be designed to be more empathic (shout out to Lev Tankelevitch, Sean Rintel, Javier Hernandez, Jina Suh, Mary Gray and Vedant Swain).

My last two projects in Microsoft were looking at how to make Generative AI more equitable and inclusive for the Global Majority. The brilliant Jacki O’Neill was a partner on both projects. Africa Health Stories is investigating how diabetes patients in South Africa can use text to image GenAI to help people make healthier lifestyle choices. Another project is looking at how GenAI can help small holder farmers in India and Kenya increase yields. Tanuja Ganu and Akshay Nambi brought their technical excellence to bear on the agriculture project. (See a great short talk about this from Jacki here.)

A farewell like this must also recognize a few other colleagues who made my Microsoft journey feel like a truly unique and wonderful experience. Steven Drucker and Emre Kiciman were always available for brainstorming and partnering. Matt Corwine and Katie Zoller were always there to help figure out creative and clever ways to amplify the stories that came out of this exciting research. Clare Morgan always brought joy and laughter to the team not to mention a mastery of patience and persistence. Dan Fay was always there to answer my “how does this work” questions. Sean Kuno and Hassan Teimoori are, in addition to being wonderful colleagues and humans, people I’m proud to call friends. Despite the long list of names, I would be remiss if I didn’t give one more big shout out to Evelyne Viegas, my manager for the past many years. Evelyne always provided me with the right advice and freedom to find my own success at Microsoft. It wouldn’t have been nearly as fulfilling or as much fun with any other manager.

Looking back what is crystal clear is how much people matter. Work is important – and I’m proud of my contributions – and being able to work with great people is the real secret sauce to a fulfilling career.  I’m lucky enough to say I had extra sauce in my time at Microsoft.

Now if you’re really still reading, you’ve reached the end. And if you’re not, then if nothing else I can use this to help explain to my kids, Sam and Jack, what my job actually was in Microsoft. That is, if they’ll actually read past the TL;DR.