An exclusive glimpse into Uber’s Data Science & Scientist internship program in Bangalore
July 10 / GlobalWe recently had the opportunity to meet with David and Javed, our data leaders at Uber in Bangalore, to discuss their commitment to investing in Data Science and Scientist Interns, and to gain insights into their experiences managing interns.
Tell us about yourselves.
David: With two decades of rich experience in data science, I’ve enjoyed a diverse career, starting as a computational biologist using algorithms to discover drugs. After a stint in sales at Intel—which, despite not being my passion, was pivotal in honing my soft skills—I pursued an MBA. My career trajectory took me from Mu Sigma, a leading data science consultancy, to managing the analytics at Swiggy, India’s largest food delivery service, and most recently at Meta in London. I relocated back to India in 2023 and was fortunate to secure a fantastic opportunity at Uber. My experience here combines the intricate challenges I faced at Swiggy with the vast global scale of Meta. Uber offers significant global roles based in India, and I am thrilled to be part of a dynamic and expanding team. Outside of the office, I’m passionate about music and play in Uber’s band, alongside Javed, our lead guitarist.
Javed: I have 20 years of work experience, with the first 5 years as a software engineer and after that, I have been into analytics and data science. My experience spans across banking and financial services, e-commerce, and now the Mobility and Delivery spaces. I lead Uber’s data science teams in Payment Compliance, Risk, Fraud, and Identity. I joined Uber in late 2020 and built this team over the last 3 years. Before Uber, I worked with eBay in Bangalore and San Jose for about 8 years. Prior to eBay, I was with HSBC bank for a couple of years. Outside of work, I like traveling for leisure, playing guitar, and listening to music (mostly hard rock and metal songs).
If you had just two minutes to explain what your team does, how would you describe it?
David: In three words: We build products!
My team, called the “Mobility and Delivery Science Team” works on all our apps to help everyone have a good experience whenever using our platform. Whether it’s a user trying to get a ride, request a bus on our app, travel from one city to another, rent a car or even get on a scooter, that’s where my team comes in to build the right product for that. We also work to ensure drivers, fleets and restaurants can succeed on our platform at all parts of their lifecycle, from sign-up and activation, to onboarding, engagement, and reactivation if necessary.
My team consists of four main parts.
- Mobility (Rides), where we work to ensure that customers are able to get the ride they want
- New Mobility Verticals, where we work on new and emerging technologies like requesting buses on the Uber app, booking rides, traveling across cities, hourly rentals, rental cars, and micro-mobility like scooters.
- Drivers, ensuring that drivers and fleets are successful on Uber at all parts of their lifecycle including activation, onboarding, early life cycle, growth, retention, engagement, and reactivation.
- Delivery, ensuring that restaurants on UberEats have a good experience and grow well.
Javed: My team works in Payments Compliance, Risk and Fraud, and Identity:
- Payment Compliance: We build products to support all regulatory requirements that a country imposes on Uber. We ensure every driver, courier and merchant in our platform is compliant and poses no risks for financial crime.
- Risk and Fraud: We proactively detect and prevent fraud to help keep our platform safe from malicious users. We also thoughtfully build products and processes to enable our good users to complete rides if they face any risk related frictions.
- Identity: We work to identify and deactivate fraud accounts at the very beginning during sign up and login. We also combat ATO (Account Take Over) problems using innovative solutions.
Describe how the data science function integrates into Uber’s overall business model.
David: We are a data-driven company, so we use data to guide decisions throughout the entire lifecycle of our products. Science teams analyze consumer behavior to identify opportunities and gaps in our products. They also help in sizing the potential of these opportunities and help with prioritization. Science teams are also responsible of building intelligence into our products, from simple things like if-then logic, to complex things like AI or ML. The team also evaluates minimum viable products to test if customers would love the product. And once a product or feature is completed, we run complex experiments to ensure it lands well, and optimize the product to increase the scale of its adoption and profitability.
Javed: The three biggest roles of data science are to monitor, measure and predict. We need to constantly monitor various business metrics to identify any problems in the business functions and track our goals/OKRs.
What gets measured gets improved: There are launches at Uber every week, and we need to be able to precisely measure the impact of each launch or change. Some of these are so minuscule in metric value (yet big enough to impact millions of our users) that only advanced experimentation science can measure them.
Prediction in making decisions: A lot of actions at Uber are taken based on probabilities of some events happening, it can be ETA of a ride, the fare of the ride, or even a user’s tendency to commit fraud. We use science to help build elaborate models with high precision and recall to achieve the right decision.
How do interns fit into your teams, and what kind of roles do they typically take on?
David: My teams have a mix of both experienced and new people. While experienced people bring maturity and knowledge, early career people bring new perspectives, energy, and the ability to look at emerging topics across the market.
Interns therefore form an important part of our team. As opposed to external hiring where we need to make a judgment call to hire in 1-2 hours, we can spend several months with our interns and the hiring call becomes much easier because of that. After an internship, it’s an easier call as to whether to hire someone or not.
Javed: Interns bring fresh perspectives and new thinking. I also expect them to challenge the status quo and ask good questions. I like them to pick some aspirational projects that may be proof of concept. Our team may not prioritize them due to constant pressure on our OKRs hence interns can be the best way to give such projects or ideas some kind of shape. Having said so, interns can also pick one of the ongoing high priority projects that are needed to achieve our immediate business goals.
What’s the learning curve like for Data Science & Scientist interns joining your teams, and how do you ensure they become productive members of the team?
David: The learning curve for scientist interns is super steep. There is so much to learn at Uber and it becomes important to pick up the skills as quickly as possible. The problems we face are complex and large. Therefore there is a need to apply innovative analytical techniques to problems. Interns have a lot of support from existing members of the team therefore despite the deep learning curve, there isn’t much to be worried about.
Javed: Uber has a very complex business model and hence I completely agree with David that the learning curve has to be very steep. In data science, if interns know SQL or some other query language it helps a lot in ramping up. However, the intern’s mentors carefully designed the internship period considering that it’s a learning experience for all.
Could you share a success story where an intern made a noticeable impact on a project or initiative?
David: One of our current interns was tasked with identifying the reasons for any trust deficit with merchants on the Uber Eats platform. He was given the task of trying to understand the quantitative feedback from surveys that we do with merchants. While he completed that work, he also felt that there would be a lot more insights when we looked at the comments and qualitative data. Instead of using the typical algorithms, he thought big and created summaries by integrating the comments with ChatGPT’s APIs. This allowed us to find key themes in the analysis. In addition, it also set the foundation for other teams to utilize the same APIs to do similar work!
Javed: Recently an intern helped clean up a lot of risk rules for higher precision. We write risk rules to stop fraudsters. These rules are basically some conditions based on facts and probabilities which when satisfied take actions on users. As situations change, these rules may become less effective and they may start impacting good users adversely. The intern carefully analyzed several rules and defined important criteria, signals, and alerts that would help change these rules or retire them as their performance degrades. This work will help us do the same things for other regions, too.
If each of you could design an ideal project for interns, what would it involve and why?
David: I give interns the same projects that existing employees would get. I treat interns like my employees but with a very long ramp-up time and maybe lesser efficiency initially. These are incredibly complex unsolved problems that have a global impact.
Javed: The projects should have good scope for open-ended thinking so that interns have the opportunity to showcase their creativity and thought processes. The ideal project must have some measurable impact and results.
What are the most unexpected challenges Data Science & Scientist interns might face at Uber, and how do you help them navigate these?
David: As they say, change is the only constant. It would be safe to assume that several things would change during the internship. A mindset that looks at change calmly as an opportunity is important. And remember that the manager and mentor are invested in helping interns succeed, so they will have the interns’ back.
Javed: Details. Solving problems at Uber may involve knowing a lot of details about a specific problem, process, or consumer behavior. It may be overwhelming sometimes to learn the required level of details and incorporate them into a solution.
Any parting words of wisdom for aspiring data scientists hoping to kickstart their careers with an internship at Uber?
David: All the best! My number one rule for success in life is that there is no substitute for hard work. So work hard. Have a good learning mindset and you will succeed in life!
Javed: Ask a lot of questions, ‘second guess’ all your numbers, and learn the central limit theorem to the level where you can explain it to a 12-year-old.
Posted by Uber
Come reimagine with us
Related articles
Most popular
How to Measure Design System at Scale
Preon: Presto Query Analysis for Intelligent and Efficient Analytics
Connecting communities: how Harrisburg University expands transportation access with Uber
Making Uber’s ExperimentEvaluation Engine 100x Faster
Products
Company