Go-Jek X DSSG

Published on: Thursday, 7 March 2019

Links to Slides can be found here:
https://www.facebook.com/download/preview/648648488888274
https://www.facebook.com/download/preview/421049648702878
https://www.facebook.com/download/preview/2290253914584368
https://www.facebook.com/download/preview/288279125405266

Synopsis
----
[Intro] Powering a SuperApp with Data Science by Maneesh Mishra (https://www.linkedin.com/in/drmaneeshmishra)

[Topic 1] How we use Machine Learning to match drivers and riders?
Speakers: Peter (https://www.linkedin.com/in/peter-richens/), Jawad (https://www.linkedin.com/in/mdjawad)

Abstract: Go-Jek, the Southeast Asian super-app, has seen explosive growth in both users and data over the past three years. Today the technology startup uses big data powered machine learning to inform decision-making in its ride-hailing, lifestyle, logistics, food delivery, and payment products. From selecting the right driver to dispatch, to dynamically setting prices, to serving food recommendations, to forecasting real-world events. Hundreds of millions of orders per month, across 18 products, are all driven by machine learning.
Our first Machine learning product for Go-Jek marketplace was a driver matching system, since then we have come a long way in improving and adapting to our growing business needs. In this talk, we will focus on Jaeger, our multi-objective machine learning allocation system.

Jawad is Data Scientist at Go-Jek, where he focuses on solving mission critical transportation and pricing problems for Southeast Asian markets. Jawad has wide-ranging experiences in Financial and Telecommunication sectors in the past. His academic work involves using simulation and data science tools to model construction workers' safety and productivity. Jawad holds a masters in Intelligent Systems Design from National University of Singapore.

Peter has been a Data Scientist at Go-Jek for two years, working on the matchmaking and pricing teams. He is a late starter in the world of technology; in previous lives, he worked for the UN and the government of Uganda. A recovering economist, Peter remains interested in causal inference and its use in combination with machine learning. He enjoys thinking long and slow about complex problems; he dislikes writing his own bio.

[Topic 2] Building complex machine learning flows for production
Speaker: Zhiling (https://www.linkedin.com/in/zhiling-chen-42764b90/)

It is often insufficient to run single machine learning models within a vacuum - we want to be able to leverage upon multiple models, perform data transformation, handle failure, among other things. Such a system involves a multiplicity of moving parts that can become extremely difficult to manage. This presentation will talk about Lasso, Go-Jek's lightweight service orchestration tool, and how it can be used to tie together the various components that make up a predictive unit, giving data scientists the latitude to build complex prediction flows while maintaining the ability to iterate quickly upon the system.
A machine learning engineer at Go-Jek, she and her colleagues work to scale ML for one of Southeast Asia's fastest growing apps. Her work aims to help Go-Jek's data scientists iterate faster, collaborate better, and serve up scalable, production ML solutions to meet the customers' needs.

[Topic 3] Burning out in Data Science
Speaker: Jireh
We address a tough topic: burning out as a data scientist. We explore reasons why data scientists experience burnout and discuss ways in which you can identify its symptoms, and discuss strategies to prevent and cure it.

Jireh is a data scientist who is passionate about helping other data practitioners succeed. At Facebook, Jireh built and maintained the AB Testing framework, Deltoid. He is currently at Go-Jek building organizational capacity for data scientists. He holds a patent for Messenger’s Sticker Search and has published in the Lancet on his favorite subject, Bayesian meta-analysis.

Produced by Engineers.SG

Help us caption & translate this video!

https://amara.org/v/nvkY/

Organization