Description
This Redbooks publication is intended to help clients with the initial steps of their AI journey on Linux on IBM zSystems. The continuously increasing interest in AI has led to a surge in availability of open source frameworks and tools, requiring enterprises to make intentional selections about using these frameworks and tools, in the context of their existing system. This publication will describe, at a high-level, the initial steps enterprises need to take in order to get started infusing AI within enterprise Linux workloads on IBM zSystems, in comparision to the traditional process of offloading AI processing to a distributed system. It will address how AI analysis and data management can benefit from the exploitation of unique IBM zSystems hardware enhancements, such as the IBM Integrated Accelerator for AI and compiler optimizations. Thus, demonstrating how enterprises can use AI in security-sensitive environments, with strict latency requirements, to implement enterprise AI solutions, such as in-transaction fraud detection and natural language processing. This IBM Redbooks Redpaper publication is intended for IT architects, developers, and data scientists.
Details
Starts 25 Sep 2023, ends 10 Nov 2023 (7 weeks) and requires 6 residents
Residency Leader: Makenzie Manna
Benefits to Resident
IBM Redbooks want technical practitioners to be recognized for their newly gained knowledge, experience, and accomplishments. Therefore, upon successful delivery of the publication produced by this IBM Redbooks residency, you will become eligible to earn an IBM Redbooks Digital Badge issued through Credly.
Additional Note: Dialogue with your management team to gain approval before applying for this residency.
Objectives
Objectives of this residency include, but are not limited to:
- Increase the adoption of AI in Linux on IBM zSystems workloads through providing examples of use case scenarios, introducing zSystems-specific hardware and technologies that are available to run AI workloads more efficiently on the platform, introducing and describing supported open source tools and frameworks, and providing initial set up information and implementation examples.
Resident Prerequistes
A basic requirement for all residents is the ability to read and clearly express concepts and procedures in common English.
Skills
Skill | Requested level |
|---|---|
| English spoken | 3 |
| Positive, self-motivated, professional | 5 |
| Teamwork and leadership | 5 |
| Software and frameworks enhanced to benefit from SMT and IBM Integrated Accelerator for AI (i.e. openblas, snapML, ONNX-MLIR, TensorFlow, and PyTorch) | 4 |
| Linux on IBM zSystems | 4 |
| Running AI analysis in Linux workloads with the IBM Integrated Accelerator for AI | 4 |
| Infusing AI into Linux workloads on IBM zSystems | 4 |
| Familiarity with popular open-source tools & frameworks used to develop, train and deploy AI models within Linux on zSystems | 4 |
| Virtualization technology (KVM, z/VM) as a platform for running AI workloads | 4 |
| Benefits of using Linux on zSystems for AI solutions | 4 |
| Options for migrating AI applications from a distributed system to zSystems (recompile, containerization) | 4 |
| Data management with Red Hat OpenShift Container Platform and IBM Cloud Pak for Data | 4 |
| Cloud Pak for Data on zSystems (for those with needs which go beyond open source offerings) | 4 |
| Sensitive data handling in AI workloads | 4 |