Data, Machine Learning, and Marketplace Optimization at Upwork (Foreword)
Overview of Data Science at Upwork
Welcome to our Upwork Data Science blog!
This is the FOREWORD of a four-part series from Upwork’s data science team and provides an overview of the team’s challenges as well as their overall vision and mission. The other installments are:
Welcome to the very first article that marks the opening of this blog we use to share learnings in our journey towards shaping the future of work!
Upwork is the world’s largest online workplace, where +5M businesses come to find and work with talents and +50M workers from +100 countries engage in flexible work arrangements.
With 30+ scientists and engineers recruited from all over the world, the Data Science team at Upwork is unique in terms of our diversity, talent density, distributed work culture and flexible work hours. Leveraging one of the largest datasets in online labor history, we work on novel solutions to tackle unique data, machine learning and marketplace optimization challenges.
The two most fundamental challenges in running online services marketplaces like Upwork are
- Promoting growth at the user level, and
- Ensuring that also at the global level, marketplace growth is healthy and sustainable.
The former is what is commonly known as user conversion, i.e., to make users happy at every step of their journey. As opposed to traditional job marketplaces, Upwork caters for the entire process from job search to hiring and job execution. User conversion at Upwork means to leverage data and optimization methods to get the users through the door, as well as to help them grow, get work done, and to establish mutually beneficial and, oftentimes, long-term relationships between clients and freelancers.
We have a very deep and contextually rich funnel, where numerous metric targets have to be accounted for and complex user behavior data can be utilized for machine learning-based conversion.
Beyond that, we recognize that making every user happy is difficult, especially as we grow our market in various job categories. More freelancers mean more competition for projects; too few freelancers with in-demand skills means clients may not find the experienced talent they’re looking for. At the global market level, we have an interesting optimization problem that requires a deep understanding of segmentation, supply and demand, and equilibrium pricing.
When we revisited the vision and mission statement of our Upwork’s Data Science team, we highlighted these two competing issues:
Our vision is to become the global data science leader behind the world’s largest and smartest services marketplace. To accomplish that, we use data and machine learning methods to innovate the core marketplace engine. This engine is so smart that seemingly, it can (1) “read our users’ mind” and help them grow into high-value core users. (2) Also, it can balance its growth to ensure a healthy and sustainable marketplace.
This is the first and broadest overview article that we put together to present a sneak peek of the various work streams we put forth to pursue this vision. Through this and future posts, we aim to:
- Discuss the opportunities and core business problems of online services marketplace operators like Upwork.
- Share insights and practical experiences in using data and machine learning to solve the core and immediate business problems of user conversion, as well as to optimize for the long-term goal of balanced and sustainable growth at the global market level.
- Connect with researchers and practitioners in the field to collaborate on unique challenges
- Provide pointers to ongoing work from individual team members and share their learnings and accomplishments at Upwork and beyond.
We aim to regularly revisit this overview to have an up-to-date presentation of our data science activities at Upwork and to add references to new technical achievements (last update: 07/072019).
About the Authors
Thanh Tran is the head of data science at Upwork, where he works with a team of 30+ scientists and engineers to innovate the core engine behind the world’s largest platform for freelancing and flexible work. As an entrepreneur and advisor of Bay Area startups, he helped built teams, raised capital for many companies and successfully shipped innovative technology solutions and end-user applications. Thanh previously served as a professor at Karlsruhe Institute of Technology (KIT) and Stanford (visiting), where he led a worldwide top research group in semantic search. He earned various awards and recognition for his academic work (Most Cited Article 5-years award, among top-5 in Semantic Search, and top-50 in Web Search per 2016 Google Scholar Global Index).
The article was reviewed and the actual work presented is done by the following data science team members: Alexander Krainov, Amro Tork, Andrei Demus, Artem Moskvin, Danylo D., Dimitris Manikis, Eva Mok, George Barelas, Giannis Koutsoubos, Hemanth Ratakonda, Igor Korsunov, Ivan Portyanko, João Vieira, Le Gu, Lei Zhang, Mikhail Baturov, Nimit Pattanasri, Pablo Celayes, Quang Hieu Vu, Roman Tkachuk, Samur Cardoso De Araujo, Sibo Lu, Siddharth Kumar, Silvestre Losada, Spyros Kapnissis, Vasily Ryazanov, Veli Bicer, Vinh Dang, Yongtao Ma, Zarko Celebic.
