Sponsored Case Study

Making Machine Learning “Magically Work” for Diverse Datasets

Let’s face it, data is messy. There are always gaps in the data and the statistical summaries stored are often non-descriptive. Why then is there a belief machine learning is magical and models will work with any data thrown its way? When building machine learning-based features for consumption by other software developers, the most difficult tasks are preparing, cleansing, categorizing, and transforming the data for use in model generation. Time series datasets come in all shapes and sizes, and each time series is its own special snowflake of seasonality, cardinality, mean, median, skewness, kurtosis, and other statistical features. This talk will take you down the technical trail of how to harness anomaly detection and event correlation, so your features will adhere to the “It Magically Works” approach to building APIs. Join us on this journey as we put our data willpower to the test.


Karlo Zatylny

Director, Architect - Tech Strategy @SolarWinds
Karlo battles the continued needs of data science and architecture at SolarWinds. Throughout his 12 years at SolarWinds, Karlo has helped engineering teams deliver solutions and help customers succeed in their ever-changing IT environments. His last four years at SolarWinds have been focused on... Read more

SolarWinds is a provider of powerful and affordable IT management software, empowering organizations to monitor and manage their IT services, networks, infrastructures, databases, and applications on-premises, in the cloud, or via hybrid models.


Wednesday May 26 / 02:10PM EDT (45 minutes)

ADD TO CALENDAR Calendar IconAdd to calendar