Achieve High-Performance Scaling for E2E Machine Learning and Data Analytics Workflows

WTH is Pandas?

Well, they’re super cute tuxedo-colored bears from China who have six toes and eat bamboo.

Also … Pandas is dataframe library that allows to you perform data manipulation in Python, and it’s great for heterogeneous data—think data science—because it provides easy-to-use APIs to manipulate and process dataframes.

(You knew we’d get there.)

But there’s an issue: When working with excessively large amounts of data or when needing high-performance, single-core Pandas becomes a bottleneck for a data practitioner’s workflow. As a result, adopting a Pandas-workflow-compatible distributed system/solution is often needed.

This webinar discusses precisely that: Intel® Distribution of Modin*.

Join software engineer Areg Melik-Adamyan for a tour of this Distribution, including:

  • An overview Modin, including its OmniSci (accelerated analytics) backend
  • How to get the best performance and scaling through Intel Distribution of Modin
  • How to efficiently run end-to-end machine learning workloads without any code changes

Get the software

  • Download the Intel® Distribution of Modin* as part of the Intel® AI Analytics Toolkit. Powered by oneAPI, the AI Kit includes 6 dev tools for accelerating data science and AI pipelines.
  • Sign up for an Intel® DevCloud for oneAPI account—a free development sandbox with access to the latest Intel® hardware and oneAPI software, including the AI Toolkit.

Other resources

Areg Melik-Adamyan, Software Engineering Manager, Intel Corporation

Areg Melik-Adamyan is an Engineering Manager and Architect with over 20 years’ experience in cross-architecture software development. Joining Intel in 2011, he currently works on next-gen data analytics, both software and its deployment across heterogeneous platforms. Areg holds a PhD in Computer Science from the Russian Academy of Sciences in Moscow and an MSc from Yerevan State University, Armenia.

For more complete information about compiler optimizations, see our Optimization Notice.