Or listen in your favorite podcast app
From joining when there were 7,000 employees to, in just four years, scaling the IT behind Uber’s current 43,000 employees, Shobhana Ahluwalia, the head of IT at Uber, was faced with the ultimate challenge. Some IT leaders dream of helping to scale their company’s IT at the rate Uber did, others may fear it. Shobhana confronted this challenge head on and now is responsible for Uber’s IT across 650 offices in 70 different countries around the globe. On this episode of IT Visionaries, Shobhana shares her insights on scaling IT at Uber, the journey automation optimization has taken over the past decades, and how leaders should think about failure.
Best advice: “Technology is best used when it is in the service of a bigger goal. And failure is not something to be lamented about when it is done in the pursuit of something big.”
- IT’s rate of change is directly correlated with the increase of data
- Scaling Uber to 43,000 employees
- Advice on how Uber views failure
Shobhana’s technical beginnings
Growing up in India, Shobana was faced with two career choices: either be a doctor or become an engineer. Shobana went with the latter because of her love of technology.
Although not all technology came easy to Shobhana at first, computer science, logical reasoning, and analogies were the components of tech that did. On the other hand, physical engineering wasn’t a particular favorite for Shobhana.
‘Growing up, things fit in my mind like a jigsaw puzzle. Technology is very logical once you understand it and understand the basics. So I loved figuring things out. I loved algorithms.”
IT and Uber
In the four years that Shobhana has been the Head of IT, Uber has grown at a rapid rate. One of Uber’s missions is to serve the cities that it is a part of and to provide services such as transportation and food delivery. Shobhana points out that Uber now has more than 650 offices around the world and Uber is used in more than 70 countries. It’s Shobhana’s responsibility to provide, “magical work experiences to all our workers, all over the world, every day,” she explains.
Uber has a varied tech stack and they work mostly in the cloud. They also have their own data centers. When you couple these computing environments together, you get what’s known as a hybrid cloud ecosystem.
‘A magical work experience is when we remove all friction in your work life. When you show up to work, everything works amazingly. You can just focus on what you want to produce. Whether that’s computers, great. Your connectivity, awesome. You have all the information that you require at your fingertips. The programs you want are working and working well. You can have amazing video conferencing going on and things like that. So basically, magical work experiences are when we’re able to remove everything external so your heart and your mind can progress without any issues and you can deliver what you want to deliver for Uber and the service to the cities.”
Being in tech for more than 25 years gives Shobhana a bird’s eye view of the IT landscape and its shifting lifecycle. Shobhana explains that IT’s rate of change is directly correlated with the increase in data being generated.
When Shobhana was working with ERP automation in the early 2000s, there was only a certain amount of information readily available to her and her team. This information was, therefore, the only information that was available to automate. Artificial intelligence and machine learning were still in their early stages of development. Fast forward a couple of decades to the present day when Uber and a number of other companies are using A.I. and ML to predict outcomes.
For example, Shobhana states, “The amount of data and the amount of information we can get is amazing. So not only have we started looking to automate this information, but we’ve also started to understand how we can use these new information sources that we have to predict the future needs of people.”
‘I remember those ERP projects, you would project 18 months to two years out into the future. And it’s very difficult for a CIO to ever go back and redo that decision because you cannot get two years back during those earlier times. The amount of money that’s spent today, it’s a whole different ball game. Our project plans are month-long if that. It’s an agile work culture. We’re always doing sprints and refiguring out what we can do and deliver in two weeks and then what can we deliver two weeks after that.”
“I would say that the technology field itself, and especially information technology, has changed so much. It’s dramatically different. The playing field has changed completely.”
Shobhana’s early days at Uber
When Shobhana joined Uber in 2015, it had a smaller footprint than it has now. At the time, Shobhana remembers Uber being a very high-flying brand, but going into the job she was unsure what the magnitude of Uber’s scale was going to be. To put it into perspective, Shobhana shares exactly what the environment was like, saying that Uber was opening 24 offices per month. By comparison, most companies will open four offices per year.
As Shobhana points out, yes she embraced the challenge of scaling Uber’s IT, but there were definitely bumps along the way. Shobhana and her team learned from failure and continued making decisions with the information she was given. An example she shares is a story of an accident that occurred four years ago during Shobhana’s early days at Uber. An engineer changed a setting on Uber’s platform that brought the entire platform down for 18 minutes. Most people’s first reaction would be to immediately fire that engineer. Not Shobhana. Instead, they went back to see what happened, what was changed, and what processes should be adjusted to make sure that kind of situation doesn’t happen again.
Some of the decisions she believes she got right were decisions made on Uber’s technical infrastructure level. She notes, “At Uber, we decided we were going to stay in our data centers rather than going to a cloud provider for our product side. On the corporate side, we embraced the clouds completely.”
“We were growing so rapidly that we were lighting up an office a day. So if a month had about 22 working days, we would probably be lighting up 24 offices around the world. Just in comparison, a normal company lights up maybe four offices a year.”
“I will say that sometimes we get caught up on how to scale or how much the scale will be. We made decisions at that point and brought in tools and systems which we thought would scale, and in some cases, we failed. To me, it’s about making the best decision you can with the amount of information you have and understanding that those decisions are changeable in the future.”