Combining AI and Open Source for Model Management and Development

This chapter explores how AI is used for model management and development, with a focus on statistical models. The importance of open source, particularly in integration with Python and R, is emphasized.

Transcript

chevron_right

Play full episode

chevron_right

Transcript

Episode notes

Guest

Joe Castle

Panelist

Richard Littauer

Show Notes

Hello and welcome to Sustain! Richard is in Portland at FOSSY, the Free and Open Source Software Yearly conference that is held by the Software Freedom Conservancy.

Today, Richard speaks with Joe Castle from SAS, a global analytics and AI company, about the fascinating world of open source software. Joe, who supports the federal team at SAS, shares the company’s journey from its roots as a statistical software company in the 1970s to its current role as an AI leader. We’ll dive into SAS’s integral involvement with open source software, how it supports and contributes to the community, and its ambitious plans for future engagement. Explore Joe’s unique insights into the motivations and sustainability of corporate open source efforts and discover how SAS balances financial incentive with authentic community engagement. Tune in for an exciting conversation about SAS’s shift towards greater open source integration and its commitment to building superior products, and ongoing discussions about making Python packages a first-class citizen in SAS. Hit download now!

[00:00:30] Joe explains his role at SAS, explaining how he supports the federal team at SAS and how it involves business development with government executives and advocating for and developing open source software.

[00:00:54] What does SAS do? Joe describes it as an analytics and AI company, and SAS clients are not just federal governments but also span industries such as banking, insurance, and more across the globe.

[00:01:51] Joe discusses how SAS uses and supports open source. Their product suite allows integration with Python, R, Lua, and JavaScript.

[00:03:33] Richard asks Joe to explain where SAS fits on the spectrum of corporate influence on open source. Joe tells us that SAS is involved in all aspects of open source usage, development, and contribution.

[00:05:36] Joe talks about SAS’s evolution from being a statistical software company to an AI company and how open source figures in their AI offerings. Developers can use Python to develop AI models using SAS’s packages and run it through their large compute engines.

[00:07:09] Joe explains his talk about the architecture of the CAS (Cloud Analytic Service), which is a Python package that allows for the sequential processing of large datasets.

[00:07:57] On the question of open source vs closed source, Joe says it depends on the context. While SAS has proprietary algorithms for model processing, developers can use their own python code to interact with these models.

[00:08:30] SAS’s primary audience includes data scientists, developers, and data engineers who have an understanding of Python and R.

[00:13:46] Richard inquires about the experiences of SAS in the realm of open source. Joe tells us they’re competitive motivations and they want to help users and capture a wider audience by signaling that they are open source friendly. He brings up the financial incentive for companies to engage with open source.

[00:15:27] Joe provides an example of a large financial customer who’s using their software for significant data processing and analysis.

[00:16:21] What’s Joe most excited to open source? He admits that some proprietary elements will remain closed due to business reasons. However, he mentions that there are ongoing discussions about making Python packages a first-class citizen.

[00:19:06] Find out you can follow Joe and SAS developer stuff online.