AI-driven insights: Uday Kiran Medisetty reveals Uber's platform evolution

AI

In the world of AI-driven innovation, few narratives captivate as profoundly as the evolution of Uber’s groundbreaking platform.

As the digital landscape continues to shift and adapt to the introduction of artificial intelligence, insights from industry insiders such as Uday Kiran Medisetty (pictured), distinguished engineer at Uber Technologies Inc., offer a tantalizing glimpse into the inner workings of the tech giant.

“We are trying to navigate the physical world, and there’s a lot of complexity the physical world brings,” Medisetty said. “We need to model in the differences in various business lines, various geographies, how different cities work, how people work in the physical world into our systems.”

Medisetty spoke with theCUBE Research analysts John Furrier, Dave Vellante and George Gilbert at the “Supercloud 6: AI Innovation” event, during an exclusive broadcast on theCUBE, SiliconANGLE Media’s livestreaming studio. They discussed the inner workings of Uber’s platform and the pivotal role of AI in driving its success.

Switching to an AI-driven model

With a career at Uber spanning almost 10 years, Medisetty recounted the challenges of envisioning a platform capable of navigating the complexities of the physical world while delivering a seamless user experience.

In developing its model, Uber needed to model various business lines, geographies and user behaviors. This led to a significant leap of faith in 2018-2019, culminating in a two-year rewrite of the core ordering system using cutting-edge new SQL technologies. The result? A robust platform capable of handling online orders and driver sessions at internet scale, powered by a new SQL spanner system, according to Medisetty.

“We cannot always keep rewriting it every one, two years,” he said. “We have to have a pragmatic choice on whether it makes the right sense for the right time. We also have to think longer term on if we were to think for the next five years, does this architecture handle the kind of scale that we are predicting for the next five years?”

Democratizing AI

Empowering engineers across disciplines to create and deploy machine learning models seamlessly is a central theme when it comes to the democratization of AI within Uber. By democratizing the creation of these models, the company aims to optimize not only end-user experiences, but also internal processes.

“I think it’s not one single machine learning model that solves every leg of the ordering lifecycle,” Medisetty said. “For us, what we try to focus on is democratize creation of machine learning models within the company.”

For example, AI-driven recommendations on the Uber Eats app optimize conversion rates, catering to both external and internal stakeholders, Medisetty explained. This approach underscores Uber’s commitment to harnessing AI to enhance every phase of the engineering lifecycle.

“How do you optimize the life of engineers with AI?” he asked. “We kind of look at every life, every phase of the general lifecycle and how we optimize that with AI.”

AI will be a transformative force capable of empowering businesses of all sizes to create Uber-like experiences, according to Medisetty. AI has the potential to streamline processes and democratize access to advanced technologies. By leveraging proprietary models and data while ensuring privacy and security, businesses can unlock new opportunities for growth.

“I think we are also in the early phase of building a knowledge graph with Uber business semantics that we can leverage to build the right LLM applications,” Medisetty said. “But in traditional machine learning models … we have a very strong data lake with really good quality data pipelines that we can use to build features that engineers use to build machine learning models.”

Here’s the complete video interview, part of SiliconANGLE’s and theCUBE Research’s coverage of the “Supercloud 6: AI Innovation” event:

Photo: SiliconANGLE