If someone had asked me to point to Milwaukee on a map I could not have done it. But on my twenty-first birthday, in 1988, I flew from New Delhi to Chicago O’Hare Airport. From there a friend drove me to campus and dropped me off. What I remember was the quiet. Everything was quiet. Milwaukee was just stunning, pristine. I thought, god, this place is heaven on earth. It was summer. It was beautiful, and my life in the United States was just beginning.
Summer became winter and the cold of Wisconsin is something to behold if you’ve come from southern India. I was a smoker at the time and all smokers had to stand outside. There were a number of us from various parts of the world. The Indian students couldn’t stand the cold so we quit smoking. Then my Chinese friends quit. But the Russians were unaffected by winter’s chill, and they just kept on puffing away.
Sure, I would get homesick, like any kid, but America could not have been more welcoming. I don’t think my story would be possible anywhere else, and I am proud today to call myself an American citizen. Looking back, though, I suppose my story may sound a little programmatic. The son of an Indian civil servant studies hard, gets an engineering degree, immigrates to the United States, and makes it in tech. But it wasn’t that simple. Unlike the stereotype, I was actually not academically that great. I didn’t go to the elite Indian Institutes of Technology (IITs) that have become synonymous with building Silicon Valley. Only in America would someone like me get the chance to prove himself rather than be typecast based on the school I attended. I suppose that was true for earlier waves of immigration as well and will be just as true for new generations of immigrants.
Like many others, it was my great fortune to benefit from the convergence of several tectonic movements: India’s independence from British rule, the American civil rights movement, which changed immigration policy in the United States, and the global tech boom. Indian independence led to large investments in education for Indian citizens like me. In the United States, the 1965 Immigration and Naturalization Act abolished the nation-of-origin quota and made it possible for skilled workers to come to America and contribute. Before that, only about a hundred Indians were allowed to immigrate each year. Writing for The New York Times on the fiftieth anniversary of the immigration act, historian Ted Widmer noted that nearly 59 million people came to the United States as a result of the act. But the influx was not unrestrained. The act created preferences for those with technical training and those with family members already in the States. Unknowingly, I was the recipient of this great gift. These movements enabled me to show up in the United States with software skills just before the tech boom of the 1990s. Talk about hitting the lottery.
During the first semester at Wisconsin, I took image processing, a computer architecture class, and LISP, one of the oldest computer programming languages. The first set of assignments were just huge programming projects. I’d written a little bit of code but I was not a proficient coder by any stretch. I know the stereotype in America is that the Indians who immigrate are born to code, but we all start somewhere. The assignments were, basically, here it is, go write a bunch of code. It was tough and I had to pick it up quickly. Once I did, it was awesome. I understood pretty early on that the microcomputer was going to shape the world. Initially I thought it might be all about building chips. Most of my college friends all went on to specialize in chip design and work at places with real impact like Mentor Graphics, Synopsys, and Juniper.
I became particularly interested in a theoretical aspect of computer science that was, at its heart, designed to make fast decisions in an atmosphere of great uncertainty and finite time. My focus was a computer science puzzle known as graph coloring. No, I wasn’t coloring graphs with crayons. Graph coloring is part of computational complexity theory in which you must assign labels, traditionally called colors, to elements of a graph within certain constraints. Think of it this way: Imagine coloring the U.S. map so that no state sharing a common border receives the same color. What is the minimal number of colors you would need to accomplish this task? My master’s thesis was about developing the best heuristics to accomplish complex graph coloring in nondeterministic polynomial time, or NP-complete. In other words, how can I solve a problem that has limitless possibilities in a way that is fast and good but not always optimal? Do we solve this as best we can right now, or work forever for the best solution?
Theoretical computer science really grabbed me because it showed the limits to what today’s computers can do. It led me to become fascinated by mathematicians and computer scientists John Von Neumann and Alan Turing, and by quantum computing, which I will write about later as we look ahead to artificial intelligence and machine learning. And, if you think about it, this was great training for a CEO—nimbly managing within constraints.
I completed my master’s in computer science at Wisconsin and even managed to work for what Microsoft would now call an independent software vendor (ISV). I was building apps for Oracle databases while finishing my master’s thesis. I was good at relational algebra and became proficient with databases and structured query language (SQL) programming. This was the era where technology was changing from character or text mode on UNIX workstations to graphical user interfaces like Windows. It was early 1990 and I didn’t even really think about Microsoft at that time because we never used PCs. My focus was on more powerful workstations.
In fact, I left Milwaukee in 1990 for my first job in Silicon Valley at Sun Microsystems. Sun was the king of workstations, a market Microsoft had in its crosshairs. Sun had an amazing collection of talent, including its founders Scott McNealy and Bill Joy, as well as James Gosling, the inventor of Java, and Eric Schmidt, our VP for software development who went on to run Novell and then Google.
My two years at Sun were a time of great transition in the computer business as Sun looked longingly at Microsoft’s Windows graphical user interface, and Microsoft looked longingly at Sun’s beautiful, powerful 32-bit workstations and operating systems. Again, I happened to be at the right place at the right time. Sun asked me to work on desktop software like their email tool. I was later sent to Cambridge, Massachusetts, to work for several months with Lotus to port their spreadsheet software to Sun workstations. Then I started to notice something alarming. Every couple of months, Sun wanted to adopt a new graphical user interface (GUI) strategy. That meant I had to rework my programs constantly, and their explanations made less and less sense. I realized that despite its phenomenal leadership and capability, it had a hard time building and sticking with a cogent software strategy.
By 1992, I was again at a crossroads in my life. I wanted to work on software that would change the world. I also wanted to return to graduate school for my MBA. And I missed Anu, whom I intended to marry and bring to the United States. She was finishing her degree in architecture back in Manipal, and we began to plan for her to join me in America.
Like all the times before, there was no master plan, but a call from Redmond, Washington, one afternoon would create a new, unexpected opportunity. It was time to hit refresh again.
* * *
On a cool, November day in the Pacific Northwest, I first set foot on the Microsoft campus and entered an unremarkable corporate office unimaginatively named Building 22. Shrouded by towering Douglas firs, it remains even today barely visible from the adjacent state route 520, known for its floating bridge connecting Seattle to Redmond. The year was 1992. Microsoft’s stock was just beginning an epic rise, though its founders, Bill Gates and Paul Allen, could