James Gan, Le Yi and Kelvin Heng (Gojek) – Gojek’s Causal Inference Based Spending Engine & A Machine Learning Optimal Pricing
James Gan, Le Yi and Kelvin Heng (Gojek) – Gojek’s Causal Inference Based Spending Engine & A Machine Learning Optimal Pricing
Abstract
A Machine Learning Optimal Pricing
Causal Inference is widely used in Gojek to provide targeted promotions for our customers. The goal is to identify the right promotions to give to a target population to maximize business objectives within certain constraints. In this presentation, we will cover the applications of SOTA algorithm, HICI and GANITE to the Demand Generation Team in Gojek.
Gojek’s Causal Inference Based Spending Engine
Finding the optimal price for rides can be a difficult task when you only get one chance to show users a price. In this talk we’ll look at the general theory behind two approaches we have used at Gojek – contextual bandits and causal inference models.
Speakers’ Profiles
Mr. James Gan
Data Scientist
GOJEK
James is a Data Scientist working in Gojek’s Pricing team.
He graduated from the University of Oxford (Master’s in Mathematics). James has worked in a variety of sectors from finance to cosmetics and moved to Singapore in March 2020, working at Gojek ever since. He enjoys the technical challenge of his work as well translating this into business impact. Other challenges he enjoys are watching Arsenal football team, playing sports and learning Mandarin.
Ms. Le Yi
Data Scientist
GOJEK
Le Yi is currently a Data Scientist working in Gojek’s Pricing team.
After graduating from NTU ( Bachelor’s in Mathematics and Economics) and working on the High Speed Rail project in LTA, she pivoted and completed a Master’s degree in Analytics before joining Gojek in 2018. Besides working on pricing problems, she was also involved in building Gojek’s matchmaking engine. In her spare time, Le Yi enjoys rock climbing, hiking and longboarding.
Mr. Kelvin Heng
Data Scientist
GOJEK
Kelvin is currently a Data Scientist working in Gojek’s Demand Generation team.
He graduated from Georgia Institute of Technology (MSc in Computer Science) and NUS (Bachelor in Quantitative Finance). His industry experience covers a broad spectrum, ranging across the applications of Data Science to Finance, Fashion, Healthcare and Ecommerce. He has deep expertise in Causal inference, images, NLP and recommendation. Together with his team, he developed Gojek’s first multi-objective causal optimisation based spending engine. During his free time, Kelvin also enjoys teaching Data Science at bootcamps to groom the next generation of tech talents.
Gallery
