Shivakumar Gopalakrishnan

Well, can you hear me now?

I'm from Varian. We are a startup, one of the first startups in Silicon Valley, started in 1948. Long time back.

And we are a company that you don't want to know. We deal with cancer care, and our product is called 360 Oncology. And what we want to do is make data-driven decisions and engage the patients meaningfully, and also enable the cancer care team. You might think of us as the ops people, and this is the requisite stack slide. I was told that I should show one of these in these kinds of conferences.

But the main thing is that our product deals with the patient at the lowest point in life, right? We promised some crying. I think this is your crying point.

What we wanted to do was, how do we solve that problem? We wanted to enable a learning culture, kind of growth mindset. And what this means to me is I always make mistakes, and my boss doesn't necessarily agree with me, but he's a cool guy. I hope this is recorded.

And this is one of my mistakes. I do that often.

On a serious note, though, computerization of medicine has been a serious failure. On the left, you see the hope: it'll be efficient. You will be on the--all the data at your fingertips. But what we did was all the doctors have become form fillers.

And if you think about it, medicine has been innovative. Clinical trials, a long time back, it's also known as the scientific method, and it has been around for a long time.

And what we wanted to do, ideally, what we would like to do, is have developers do multiple experiments. They can go back in time, Star Trek reference. I said that was another requirement, they said. Then this also provides them a safe environment, right?

Obviously, the patient care side, we don't want to mess with it. We take our security and PHI very seriously. So we wanted a safe environment for this one. And developers are very picky people. They also apparently want to have fun doing this. And you can't satisfy these people.

We were really thinking, "How can we even do this?" And as usual, DevOps seems to be the solution for everything. So we said that, "Why can't we think of DevOps issues?" And also we get free development from the developers.

Computers are like people, too. They act weird, whether they are pets or cattle. They have their moods, and they act crazy sometimes. And I'm not even talking about software. Just the hardware is crazy enough.

And once we get the developers to solve that problem, then we show them our tool stack. It's like, whoa.

And we want to do machine learning. Every time there's a new tool stack, you know, TensorFlow, and so how shall we do it, right?

And I shall pause, I think, to catch my breath. And I've never talked to so many people before. I'll enjoy the favor of the moment for the next five seconds.

The solution for us was to do Google Sprint. The key thing was that it enabled us to involve the business people. We involved the product managers, service people. They get to vent out in all their problems, and the solution that we came out with, we'll do a chatbot.

It'll tell us your status of the deployment, how the customer performance is doing, anything that you can ask. As long as it fits in with AI rule, it'll answer.

And of course, developers being developers, they threw the kitchen sink at it. They threw Alexa, Lambda, serverless computing, event-based integration, AI for Watson, and mobile, and all the stuff that we wanted to do.

And they were able to play with all these technologies like microservices, SSO, composable UIs. These are really important for the future of our business if we want to deliver innovative applications faster to the customer.

At the end of the day, we want our mission to be releasing impactful changes to our customers, and happy developers result in high-performing organizations.

Thank you very much.