The AI Engine Behind Uber’s Mobility Operations

In an exclusive conversation with Dataquest, Manikandan Thangarathnam, Senior Director of Mobility and Platforms Engineering at Uber, delves into the company’s technology-first approach. He explains how AI fine-tunes route planning, optimizes fleet operations, and even powers Uber’s dynamic pricing models. 

From ensuring smooth third-party fleet integration to addressing India’s diverse mobility needs, Thangarathnam offers a deep dive into Uber’s evolving tech stack and its ambitious roadmap for the future.

With Uber handling over 25 million trips per day, how does the company ensure data accuracy, operational efficiency, and regional adaptability—while preparing for an AI-driven mobility future?

Let’s find out…..

How is Uber leveraging AI and machine learning to enhance core operations, such as efficient route planning, dynamic pricing, and fraud detection, while also improving user experiences through personalized support and product recommendations?

At Uber, we integrate AI and machine learning to enhance operations across several processes and flows. Route planning is one of the many such flows where we leverage industry-leading AI algorithms to analyze traffic and use data for optimization. In the case of High Capacity Vehicles (Uber Shuttle) and shared rides, this ensures minimal walking distances and maximum vehicle utilization, while dynamic models adjust prices in real-time based on demand and external factors.

Our AI models for HCV, for instance, are trained to optimise experience based on four key parameters - seat utilization, rider comfort, vehicle utilization, driver utilization.  

Additionally, our machine learning systems enhance user experience by tailoring the Uber app to each rider’s needs by providing personalized product recommendations based on preferences and availability.

Could you share insights into how Uber drives data excellence in its operations and the role of emerging technologies like AI and IoT in achieving this?

We drive data excellence by harnessing AI algorithms for anomaly detection, effectively pinpointing errors within the vast datasets from the 10,000 cities we operate in. This approach allows us to quickly identify and address discrepancies at a micro level, ensuring the integrity and accuracy of our operational data. Given the scale we operate at, certain configuration changes might have the potential to cause a matching error in a smaller city such as Raipur, for example. The problem statement for our teams is to code in a way where these anomalies, even at the smallest scale, are figured out and corrected within minutes.  

How does Uber's platform-based software adapt to regional variations without disrupting operations, and what role does technology play in this adaptability, particularly for fleet management?

Uber’s platform-based software is designed to seamlessly adapt to the diverse compliance and regulatory requirements of each country while maintaining uniformity in operations across the cities we serve. Our centralized codebase allows local operations teams to implement region-specific modifications easily, ensuring that these adjustments meet local laws and cultural nuances without compromising the integrity of the global system.

This local configurability, supported by robust technology, is crucial for efficient fleet management and operational consistency. The real challenge lies in ensuring that these localized changes harmonize with our global standards and do not disrupt the overall user experience. Technology, particularly advanced analytics and machine learning, plays a pivotal role in evaluating the effectiveness of these regional modifications across our vast network, enabling us to maintain a high-quality, cohesive service experience worldwide.

How does Uber’s technology platform enable real-time optimization and scalability of its fleet operations, especially with the inclusion of third-party vehicles, while ensuring seamless regional adaptability?

Uber’s technology platform offers a comprehensive end-to-end fleet software solution that enhances real-time optimization and scalability of fleet operations, including the integration of third-party vehicles. Unlocking supply is one of the key business needs and our fleet technology supports it in a big way.

By providing detailed data on revenue and utilization, along with real-time updates on trip status and driver activity, our platform ensures that fleet operators are always making informed decisions. We also extend API-based integrations for larger fleets to seamlessly connect their systems with ours.

Currently supporting over 20,000 fleet partners managing fleets as large as 10,000 cars, our software is optimized to highlight key performance indicators such as underutilization, upcoming incentives, and driver performance, ensuring operational efficiency and regional adaptability. The tech for fleets has been built by our tech centers in India.  

What advancements in Uber's tech stack support the integration of third-party fleets into its ecosystem? How does the platform tackle challenges like data consistency, driver performance tracking, and vehicle compliance across diverse operators?

For fleet operators who prefer to use their own applications, our platform supports robust third-party (3P) integration, allowing trip requests to be routed directly to their respective apps. This integration has been successfully implemented with partners such as New York Taxi in the US and Shadowfax in India. We provide a standardized API that third-party operators need to integrate with. 

This ensures that data shared between Uber's platform and third-party apps is consistent and conforms to the same standards. It helps in synchronizing information like vehicle availability, trip details, and driver status in real-time.

We have agreements in place with all the 3P operators that require them to meet all conditions on safety and regulatory requirements, including necessary documentation for drivers and vehicles, along with background checks. These checks ensure that vehicles meet Uber’s safety standards and local regulatory requirements. These advancements ensure that Uber's platform can adapt to varied operational needs while maintaining a cohesive and efficient service network.

How does Uber ensure scalability and reliability when onboarding third-party taxi operators into its platform? Are there any technology-driven measures to ensure smooth integration across operators with varying levels of digital readiness?

We provide clear guidelines for third-party partners on essential features required for integration, ensuring a baseline functionality that aligns with Uber's standards. Our technology is tailored to accommodate specific needs, offering bespoke metrics for each partner and enabling on-demand features such as multi-dispatch, which sends trip requests to multiple drivers simultaneously. This allows those fleet partners to utilise their inventory more effectively. 

This approach also optimizes the rider experience by improving the likelihood of faster service, even across operators with varying levels of technological advancement. These processes ensure scalability without compromising on reliability on the platform.

India has diverse urban and rural mobility needs. What localized technological innovations is Uber developing to cater specifically to the Indian market, including smaller cities and underserved areas?

Our commitment to address India's diverse mobility needs through localized technological innovations remains a top priority as we innovate for the future, especially in smaller cities and underserved areas. We've localized the Uber app in multiple regional languages in the country and introduced India-specific features such as cash payments to accommodate the prevalent transaction mode. Recognizing local nuances, we enable operations in local currencies across cities, despite the complex technical challenges involved. 

Additionally, we enabled in-app phone calls, acknowledging the preference among some riders to communicate with their driver before the ride. These adaptations ensure that our app is finely tuned to meet the specific needs and preferences of local users, enhancing their overall experience with Uber.

What are the primary challenges Uber anticipates in implementing its 2025 technology roadmap, and how does the company plan to address them?

As we continue our organic growth, clocking over 25 million trips per day, we face new challenges in an evolving transportation ecosystem. Our key focuses include integrating EV technologies to address aspects such as downtime and charging, and expanding autonomous vehicle partnerships to streamline integration processes. 

We are committed to maintaining an open platform which facilitates smoother integration and adaptation. By simplifying and enhancing our technological infrastructure, Uber is well-prepared to address these emerging challenges and sustain our momentum in the evolving transportation landscape.

Looking beyond 2025, what emerging technologies does Uber plan to explore to maintain its competitive edge in the mobility sector?

As a pioneering tech company in the ridesharing industry, we consistently seek out emerging technologies to maintain our competitive edge. Building on a legacy of innovation established over the past decade, we continue to position ourselves as the prime destination for engineers eager to tackle significant industry challenges. By staying at the forefront of technological innovation, we aim to lead the evolution of the mobility sector, providing cutting-edge solutions that enhance urban transportation globally.