Abstract: Fiverr, the talent cloud solution for businesses of all sizes, leverages cutting-edge AI techniques to provide value in a myriad of business use-cases, both in their core product and in its periphery. From recommendation engines, via spam and fraud detection, and all the way to marketing enablement, Fiverr’s data-science team tackles critical business problems by implementing intelligent monitoring solutions designed to gain valuable insights into their AI systems.
In order to make sure their investment in AI bears fruit, Fiverr’s core data science team has long realized that they cannot stay focused only on research and must adopt a product-oriented approach to their work. By taking ownership of model performance in production, they guarantee not only exceptional performance in controlled environments (test set, A/B test, or the offline environment) but also optimal behavior and results in real-world scenarios.
Mona, a proactive observability platform, plays a pivotal role in Fiverr’s journey. It empowers data scientists to collect invaluable insights on model behavior and business outcomes, providing alerts early when AI systems underperform or misbehave, mitigating risks to business KPIs.
In this talk, Gal (Senior Data Scientist, Fiverr) and Itai (CPO, Mona) share how Fiverr utilizes advanced tools, both home-grown and bought, to bridge the gap between data science and business, empower data scientists to understand the behavior of their models in production and make sure their AI solutions bring the value they’re expected to deliver.
To do this, Gal and Itai will discuss concepts and principles in product-oriented data science and then ground these concepts with real-world examples from Fiverr’s experience.
Key topics discussed are:
- The creation of a machine learning platform that abstracts productionization from the data scientists, enabling rapid development cycles, as well as a shared format of training, inference, and monitoring pipelines
- Utilizing advanced monitoring and analytics to make sure the models perform as expected in the context of the business function they serve
- How understanding your model in production helps the data science team do better research, leading to further innovation and breakthroughs
- How taking ownership of the model in production improves the chances of successful data science projects and maximizing value delivery
Bio: Gal Naamani has been working as a data scientist for 4 years, with the past 3 years being at Fiverr. As the Senior Data Scientist, Gal works closely with developers, analysts, product managers, and business owners on growth opportunities and new ideas, from research to production. Gal currently has leading roles in projects that are focused around search engine ranking, promoted ads, online bidding optimization, exploration-exploitation problems, monitoring, and more.