Alum spotlight: Vikas Goel

Vikas Goel ('05) has patented six of his innovations for using modeling and data to make better decisions when there is uncertainty in the data. These patents include application to production optimization, long term investment planning, and supply chain optimization in the oil and gas industry.

When Goel saw the growth of data science-based decision-making, particularly when fueled by artificial intelligence, he transitioned his career from the oil and gas industry to e-commerce. "I think it is a really good playground for these technologies," he says. The e-commerce industry offers him a different pace and way of doing work compared to the more mature oil and gas industry.

Goel is senior director for data science at Flipkart, one of the largest e-commerce marketplaces in India. Working on the supply chain side, Goel leads a team that builds models to optimize the post-order delivery process. Flipkart delivers millions of orders each day, and Goel's team uses data to reduce delivery costs while increasing delivery speed and reliability.

E-commerce in India is growing and evolving rapidly, meaning businesses face more variability than in a mature market. Customer expectations vary, and their behaviors are changing. "There are customers who want deliveries very quickly and are willing to pay for that. Other customers want to control their costs and are comfortable with slower speed," says Goel.

In response to the changing marketplace, Goel's team built an advanced supply chain planning platform in only three years. "The platform changed the way our supply chain planning works across the entire planning space: inventory planning, capacity planning, and network planning," says Goel. His team was one of six finalists for the 2025 Franz Edelman Award, the pinnacle of achievement in the application of analytics and operations research.

In addition to core supply chain optimization, Goel works on problems in fintech, fraud detection, and geocoding, a challenge posed by India's non-standardized addresses. "When a customer places an order, they enter a text address. We have to figure out where that location is before we can begin to figure out how we will deliver," says Goel.

His experience as a chemical engineering Ph.D. student at Carnegie Mellon uniquely prepared him with both the skills and the tenacity to solve open-ended problems. As a Ph.D. student, Goel spent months working on a problem of making investment decisions in the oil and gas industry before he realized he was on the wrong path and had to start again. "At that point in my life, that was my biggest experience of falling down and standing back up," says Goel. "It helped me build resilience and prepared me for taking risks in my career."

The research challenge lay in the interplay between planning decisions and the uncertainty in data. Oil or gas field developers have to make significant up-front investment decisions based on uncertain estimates for the amount of resource in the field. This problem was unique in that the choice of decisions could influence what uncertainty is reduced and help make better decisions in the future, an aspect that had not been studied before. Goel pivoted his dissertation to explore this dependency between decisions and uncertainty. The approach that Goel developed with his advisor, Ignacio Grossmann, has since been applied in a range of other areas.

Goel emphasizes the broad base of coursework in chemical engineering and business from which he built technical expertise in optimization while at CMU. The ability to bring those areas together has helped him succeed in cross-functional work.

"In the times that we are in, when AI is really blossoming, students at CMU have a great opportunity to build a broad technical background spanning chemical engineering, computer science, and operations research and management," he says.