Data Science and Analytics (DSA) at Moloco

Data Science and Analytics (DSA) at Moloco

About Moloco

Moloco’s goal is to make the digital economy more equitable and profitable by delivering advanced machine learning to companies of all sizes. With Moloco’s machine learning platform for growth and performance, every app publisher and online retailer can now unlock the value of their unique, first-party data for user acquisition, retention, and monetization campaigns.?

Founded in 2013, our advanced machine learning engine powers our product portfolio. Moloco Cloud DSP enables performance marketers to quickly scale user acquisition and achieve greater lifetime value through battle-tested prediction models. Moloco Retail Media Platform enables online retailers and marketplaces to establish their own performance ads business.?

Our technology is best in class; we received the?SMARTIES X silver award?for Machine Learning and AI, and Moloco was named the?Cross-Industry Winner?for Google Cloud Customer Awards.?

The company is in hyper-growth mode and we ranked #95 in the Inc. 5000 fastest-growing private companies for 2022. We ranked #91 among?Deloitte’s 2021 Fast 500?and have been certified by 91% of the company via?Great Places to Work.?

ML, Data, and Algorithm Challenges

At Moloco, we solve extremely large and complex problems. In the DSP product, on behalf of advertisers we publish advertisements on quality ad spaces at low prices, while maximizing the effect of advertisements using our ML technologies in the real-time bidding (RTB) ecosystem (Figure).

No alt text provided for this image
Figure. Real-time Bidding (RTB) Ecosystem

The scale of the problem is immense:

  • 60B+?Impressions served monthly
  • 5.5M+ Predictions per second
  • 2B+ Devices (users)? reached monthly
  • 0.5M Apps reached monthly
  • 100ms end-to-end response time limit
  • O(10^3) campaigns

The core infra, ML, and algorithmic challenges include:

  1. Conversion Prediction: predicting the probability that a user "convert" (install an app, in-app-purchase actions, etc) and expected value of such actions from contextual, user, and campaign features.
  2. Market Price Prediction: predicting the probability of winning the 1st-price as well as the 2nd-price auction as we change the action price
  3. Budget pacing: spending campaign budgets in an "optimal" way so budgets are fully spent over time while maximizing the camp

and many more.

Moloco Infra Eng and ML teams are hard at work solving the above challenges and data science and analytics plays critical roles in solving these challenges partnering with Biz and Eng partners.

Data Science & Analytics (DSA) in Moloco

Data Science and Analytics (DSA) is the global team of data scientists, data analysts, and analytics engineers. The 3 job families bring unique superpower to the complex, data-intensive problem space:

  • Data Scientists (DS): you will derive performance improvement and cost efficiency in our product through a deep understanding of the ML and infra system, and provide a data driven insight and scientific solution.
  • Data Analysts (DA): you will provide data products, e.g. raw data, campaign facts, growth insights to help customers improve campaign performance and maximize their business goals with Moloco.
  • Analytics Engineers (AE): you will build data-warehouse resources and tooling to enable data consumers.

DSA team's vision is to "Deliver high quality data products at scale to DSP while supporting success of new product (RMP & PAS)" and is organized as following:

No alt text provided for this image
Figure. DSA Teams and Biz / Eng / Prod partners

The DS Team is functionally sub-divided into the following groups:

  • Infra DS : focus on development projects for our bid-processing infra
  • Cloud DS : focus on development projects for our DSP Cloud, to empower internal and external DSP users to manage their campaigns successfully through our data-based guides, estimators, and recommendations?
  • Experimentation DS : focus on data science work to enable scalable A/B testing infra, develop high quality metrics, support incrementality tests, research and develop advanced methods for nonparametric (Jackknife) testing, variance reduction, and multiple comparisons, etc.
  • Retail Media Platform (RMP) DSA: Moloco Retail Media Platform enables online retailers and marketplaces to establish their own performance ads business.?(Think "Sponsored Ads" in Amazon.) RMP DSs and DAs work with recommendation and search ranking ML team

The global DA Team is regionally organized as follows, since they need to work in close proximity with Biz counterparts:

  • US DA (in Redwood City and Seattle)
  • EMEA DA (in London)
  • APAC DA (in Singapore and Seoul)


Between DS and DA teams, we do the following activities to help improve our product and services, while bringing operational excellence and high quality insights for better decision makings.

No alt text provided for this image
What DS and DAs do at Moloco DSA


Why Work in DSA?

  • Culture - DSA inherits excellent Moloco Values, namely (i) Uncapped Growth Mindset, (ii) Create Real Value,?(iii) Go Further Together, and (iv) Lead with Humility.
  • People - DSA has numerous (50+) smart, passionate, humble, and hard-working team members. Many are top school graduates and Ex-FAANG.
  • Growth - The company and team is small enough to have ample room for impact. Company is in hyper-growth mode and DSA team is growing as well to support growing Biz/Eng/Prod needs.
  • Stimulating problems - Solve hard technical, operational, and business-related challenges. Opportunities that come with multi-million QPS, peta-scale data, billions of users, numerous online and offline ML models, and real-world customer data to analyze and investigate

We're Hiring!

See https://boards.greenhouse.io/moloco for the currently open jobs at DSA and across Moloco.

Robert M. Dayton

MBA, Engineer | Enterprise AI | Advanced Analytics | GTM Strategy | World's First Arbor Essbase Post-Sales Consultant

1 年

Thank you for sharing Jaimie!

回复
Douglas J.

Principal AI Scientist | ex-Google | GenAI Leader | Innovation & Performance in AI Applications

2 年

Hi Jaimyoung, this number looks so huge: "21B+ Devices (users)?reached monthly"; shouldn't it be smaller than the number below: "15M Apps reached monthly" (as the ad can reach multiple Apps for the same user/device)?

要查看或添加评论,请登录

社区洞察