Session: Fast iteration and quick ideas validation in the Machine Learning world.
Have you ever invested long days in building a feature or creating a Machine Learning Model that never ended up impacting the user behavior? Or as a Data Scientist, have you ever struggled with identifying the most effective parameters for your model?
In our company we observed many situations like this, until the teams started applying the Experimentation approach to their product development. This concept is commonly known as AB testing, where user behavior is compared between groups using the changed and unchanged version of the product. Not only did it allow them to quickly validate the correctness of their ideas but also enabled Data Science teams to iterate over their models to establish the best parameters.
In the talk, I would like to share with you why Experimentation became extremely important for my company, how it is used in the AI projects and what benefits it offers.
Bio
I am a strong believer in the Data-Driven approach and currently I am leading the team responsible for driving the Experimentation culture within the company. As a team, we are building an internal tool for running AB tests, we spread the spirit fo and belief in Experimentation Importance and we engage and motivate our internal users to apply the Data-Driven approach into their day to day work.
Personally, I am an ambitious, devoted, loud, zingy team player, who started a career not more than 5 years ago, raising from Monetization Analyst, through Product Analyst to Product Manager.