Engineering a Million-Mile Journey with Uber ATG

Poornima Kaniarasu is a data platform engineer with Uber’s Advanced Technologies Group (ATG). In this article, she shares how her background in robotics led her to Uber Engineering, where she works on the technologies powering Uber’s self-driving vehicles.

“Are you sure this is the place?” My Uber driver asked as he pulled up to my destination.

I was pretty sure I inputted in the right address to my Uber app, but this definitely did not look like a corporate building. A few of my friends from grad school had recently started working on autonomous cars for Uber at their Pittsburgh office, and while I was not really sure if I wanted to move to Pennsylvania, I was curious about the position so I decided to take the call.

Hesitantly, I got out of the car and entered the Uber ATG office, and a receptionist confirmed that I was indeed in the right place.

Once I entered the building, my concerns, quite literally, flew out the door. The ATG headquarters was reminiscent of my grad school days: people sitting around open tables, throwing ideas at each other and whiteboarding, scientific scribbles all over the walls. The atmosphere was warm and welcoming, and the people were passionate and curious. As a recent college grad, I found myself yearning to work here. Everyone I spoke to was so enthusiastic about what they were doing. I got the impression that people really wanted to work here, as their goals aligned with Uber’s objectives.  

After my interviews, my mind was completely made up. At this point in my young career, joining a fast-growing, rapidly innovating company was exactly what I wanted, and working for Uber ATG would give me the opportunity to tackle technical challenges from the ground up that were shaping the future of transportation. When I received my offer letter, I could not accept fast enough. Nearly two years later, I can say with certainty that Uber ATG was the right decision for me.

Since I was a young girl, my mom has been a constant inspiration and guiding force behind my decision to focus my academic, and later, professional pursuits on STEM. After graduating from university in the 1980s, she pursued a career as an electrical engineer, a remarkable feat for a girl who grew up in the conservative society of southern Tamil Nadu, India. Not surprisingly, her professional ambitions were short-lived due to the social constraints of her community.

But that did little to deter her passion for math and science. For example, she did not hesitate to take apart household electronics if something was wrong with them. As a kid, I would sit by her side and watch with fascination as she debugged the issue. Decades later and encouraged by her example, I set out to accomplish what she had been unable to, and throughout my engineering career, she has always been my pillar of support.

When I began my undergraduate degree in 2006 at PSG College of Technology in Coimbatore, personal computers were starting to catch on in India. I purchased my first computer during my sophomore year, and was introduced to robotics and embedded systems the year after. In fact, my first robotics project was building a mini robot for an inter-university competition in which robots navigated a virtual city. Points were awarded to robots who successfully made turns, stayed on their specified route, and completed the circuit in the shortest amount of time. We hacked a toy car by adding micro-controllers and a webcam to it, which enabled it to interpret and react to traffic lights and street signs. This learning experience gave me the confidence and enthusiasm to pursue my masters in robotics at Carnegie Mellon University (CMU).

My focus at CMU was developing trust in human-robot interactions. I researched how a robot can effectively convey its intention to humans and how humans can foster more trusting relationships with machines that they regularly interact with to accomplish simple tasks. At CMU, I also served as a research intern in a lab that was developing new modules for the Braille Writing Tutor (a device that uses audio feedback to teach individuals to write braille) and researched ways to extend the service to devices such as smart phones. When I worked on the Braille Writing Tutor, I learned the value of feedback, receiving input on how to create a user-centric design from the students and teachers using it on a daily basis. Developing a technology that people actually used and hearing from them directly that it improved their lives was incredibly satisfying.

Working at Uber, I have been able to do this on a much larger scale. Building reliable, accessible transportation options that will improve lives worldwide is a powerful experience.

At ATG, I am a data platform engineer on the Machine Teaching and Interactive Learning team. We are  responsible for producing the ground truth data that is used to train various machine learning algorithms across the organization. Autonomous car perception, motion planning, and map building are some of the major consumers of the data we generate, and it can be challenging to generate this data at scale with high accuracy. Still, I find it interesting to work on developing intuitive interfaces that take advantage of the strengths of humans and algorithms, catering to their complementary skill sets.

In March 2017, Uber organized our first-ever Uber Technology Day, featuring tech talks from LadyEng members across our organization. During the event, I delivered a presentation on segmentation for object detection that gave attendees (engineers from Uber and other technology companies in the Bay Area) a behind-the-scenes look at some of the projects I was working on at ATG. My involvement with LadyEng, Uber’s women in engineering group, as part of Uber Technology Day has been one of the highlights of my time at the company. It was a nice occasion to interact with and field questions from women working in machine teaching or those who were curious about the subject.

I hope that by sharing my experiences and talking openly about the exciting work I get to do as an engineer, I can encourage more women to pursue careers in STEM. In elementary school, I was taught that Charles Babbage is the father of computer science, and it was not until grad school that I first learned about the achievements of Grace Hopper. Grace Hopper, an inventor of one of the first compilers, is a role model for many young women, myself included. Imagine how far my mother could have gone if she had role models who deftly navigated her conservative community and were able to continue their careers in engineering!

It is not hard to find inspiring women in Uber’s Engineering and my experience at Uber Technology Day exemplified this.

One of my proudest moments at ATG so far was the day we launched the first self-driving Uber pilot in Pittsburgh. Less than a year later, self-driving Ubers have driven over a million autonomous miles in the U.S.

From my first encounter with the Uber Pittsburgh office to my experience during Uber Technology Day and reaching the million-mile mark, my time at Uber has afforded me the opportunity to be a part of an ever-evolving, increasingly important industry that is shaping the future of transportation.

There is no doubt: I am definitely in the right place.

And you can be, too! If you are interested in engineering solutions to help train machine learning systems, consider applying for a role on our team.

Poornima Kaniarasu is a software engineer with Uber ATG.