Tag Archives: AI/ML

Lean In Data Science

How do we apply the principles of lean to data science and data engineering? We discuss this broadly into using AI and machine learning more generally.

This is a topic that we had discussed over the summer and wanted to come back to six months later because so much has changed and transformed in the industry. What does agile lean process control look like in an infrastructure automation platform? How can we make these very difficult and challenging components of data and data management, more agile, more lean?

I think you will get a lot out of this conversation considering our current hypercharged AI ml and LM environment.

Transcript: otter.ai/u/1ZuALgSXcPw-bIf2GO…?utm_source=copy_url
DALL-E Prompt: please create a picture of a very large truck stuck under a low bridge. please label the truck as ai and the bridge as lean

Generative Coding & DevOps Challenges

What can we expect generative AI to generate and is it going to produce good code? Today we talk about Gluecon and generative DevOps and the different concepts and capabilities around it. What impact is it going to have on developers? How do we control that?

Today’s discussion was in preparation for our session on June 13, where we’re going to group program GPT to see what type of DevOps coding skills we can prompt. We talked about the necessity of prompting in this session and covered some tips to help you think about how to be a better prompt engineer, a skill set that everybody’s going to need to have in the next months if not years.

Transcript: otter.ai/u/lsye_htH0-wksAOrqg…?utm_source=copy_url
Image: www.pexels.com/photo/man-welding…ow-frame-2965260/

Domains And Access For Metadata

What actually is used to describe the provenance and information that comes with our data? Today we discuss metadata and the governance, security, hint, domains, date that accompany data in ways that in some senses are more important than data.

How can we move, change and transform data? We had a really robust conversation about how access to data is so critical in actually understanding how data is used.

This is a topic we struggle with: figuring out the naming, how things work, and the context. More than anything else, the context makes things challenging. As you listen to this, think through how challenging it is to define data, data governance, and using data effectively.

Transcript: otter.ai/u/VE2u_jTwG_h4z9ewfI…?utm_source=copy_url
Image: www.pexels.com/photo/cheerful-bl…hoolkids-5905918/

Generative DevOps

NOTE: This is Rob’s Gluecon topic on 5/24. Save $300 if you register with speaker300 at www.gluecon.com

We dive into the question of whether or not generative AI can be used to productively change DevOps automation and the control of infrastructure.

We’ve discussed the closed loop side of using AI to manage infrastructure in the past, but this episode we really dive into the idea of creating automation and using generative AI.

Transcript: otter.ai/u/VtnznHgydT3_6QSJpk…?utm_source=copy_url
Image: www.pexels.com/photo/tossing-fri…ying-pan-6937457/

ROI from Putting Data In Context

https://soundcloud.com/user-410091210/roi-from-putting-data-in-context

If you love data and data context formats for exchanging data, you will love this conversation.

Today’s episode is a deep conversation about the potential ability to define ways in which we produce, store and share data, providing context using markup languages, and then being able to extend that. It’s a fascinating conversation about how much we could improve our use of data if we were able to provide more context about who wanted to see it and what relevance it had.

We also have some interesting conversations about data migration and how we share information.

Transcript: otter.ai/u/S9_tibMGhkoajG0bP8LaciJ32Ys
Image: www.pexels.com/photo/a-grayscale…is-hand-10839215/

Serverless vs Digital Twins

We discussed the intersection of serverless and digital twinning. These two concepts are really tightly intermingled!

We discarded the idea of a central single serverless hub managing everything; instead, we think sites would actually have a mesh of serverless, interconnected event processing and stream processing systems. This approach is much more function dependent, but really opens up a lot of interesting discussions and possibilities.

We also discussed how to manage all of this meshed, serverless subscription modeling eventing, and digital twinning.

Transcript: otter.ai/u/FWW3CwG6gMX0N06QXp9C-pZd-s8
Image: www.pexels.com/photo/two-young-g…-sweater-9532902/

Cloud2030ServerlessCloudDigital TwinsModelsAI/MLEvent ProcessingDistributed SystemsEdge

Rob’s Hot Take:

Rob Hirschfeld, CEO and co-founder of RackN and host of the Cloud 2030 roundtable discussions, shares insights from the January 13th discussion on serverless and digital twinning. He emphasizes the significant overlap between modeling data through digital twinning and building serverless systems that subscribe to events for action, transformation, and control. Hirschfeld underscores the need for a joint discussion on these topics, highlighting that a sustainable serverless system requires a reliable model, and a useful digital twin system must connect inputs and outputs to the real world. Interested listeners are invited to explore the full episode at the2030.cloud for a comprehensive discussion on these critical industry aspects.

Machine Learning in Operations

Today’s episode is about how to trust machine learning in operations. This is a really serious issue because the attraction of machine learning is strong, but does not translate into operations.

Why doesn’t it translate? Because operations is a closed loop process where we constantly get feedback and have to adapt and adjust. That makes it difficult to train models and hope that they work. This discussion gets into why that’s the case and what we can do about it.

Then we explore scenarios for machine learning and AI in operations.

Transcript: otter.ai/u/UBjf5IVnKvebTfQgW1xlneZdjOU
Photo: www.pexels.com/photo/high-angle-…of-robot-2599244/

Edge Impact of Digital Twins

We talk about Digital Twins and the Edge with Simon Crosby from Swim.AI. They are literally building digital twins in edge locations so he has a lot to share.

We work to expand and understand how Simon’s experience translates into general cases and what we’re seeing in the edge. The systems that we’re trying to build are at the intersection of models and “connectedness” of all the components for the edge.

These designs don’t fit traditional models and it is what makes edge unique. Edge is not a single application, but a connected system that going to have to emerge to make all this work together.

Transcript: otter.ai/u/-uFSclONwRhhc4QlFywiSJAIF10
Photo by Dmitriy Ganin from Pexels [ID 7538096]