Agile Mumbai 2022 - Ashwinee Singh | Agile in AI or AI in Agile?
1. Agile in AI
OR
AI in Agile
Ashwinee Singh – India Head of Business Agility
ashwinee.singh@ust.com
November 2022
Agile Mumbai 2022
www.agilemumbai.com
https://towardsdatascience.com/crisp-dm-ready-for-machine-learning-projects-2aad9172056a
This has an additional phase of "model management to incorporate model monitoring and to make necessary changes based on performance. Since there is no standard version of CRISP-DM for AI, various improved versions are available; we find this one more suitable. Except for one point, in addition to the connecting path from Model management to data preparation, there should also be a path from model management to modeling. (this is because during model management, if results are not expected, then sometimes we need to fix data and sometimes fix or change only the model or both. It is up to you if you find it suitable to incorporate into the storyline )The "data modeling" phase is replaced with "modeling" w.r.t the previous slide. This is to accommodate ML modeling. This has alternate suggested dependencies of phases. Please see if this is useful, and add/remove/change as you find it suitable. The rest of the things - DoD, DoR, roles, etc. are kept and mapped as-is basis, per slide no. 22. Data scientists will be involved in the evaluation/validation of models using accuracy and other validation metrics. They will be involved in model management. They will provide the necessary support to the ML engineer team for deployment. Hence the overlapping areas are adjusted.
AIOps is a way to automate the system with the help of ML and Big Data, MLOps is a way to standardize the process of deploying ML systems and filling the gaps between teams, to give all project stakeholders more clarity (https://neptune.ai/blog/mlops-vs-aiops-differences#:~:text=AIOps%20is%20a%20way%20to,all%20project%20stakeholders%20more%20clarity.)
MLOps and AIOps: https://www.analyticsinsight.net/what-are-mlops-and-aiops-how-do-they-differ/#:~:text=MLOps%20doesn't%20specifically%20refer,MLOps%2C%20despite%20the%20obvious%20distinctions.
MLOPs and AIOps
https://www.analyticsinsight.net/what-are-mlops-and-aiops-how-do-they-differ/#:~:text=MLOps%20doesn't%20specifically%20refer,MLOps%2C%20despite%20the%20obvious%20distinctions.
Context: MLOps for "Agile in AI" (Machine learning model deployment and management), and AIOps for "AI in agile" (AI and ML for IT operations/projects) -------- The links for reference is added in the notes section. Additional notes MLOps can be considered as DevOps for machine learning pipelines. Putting ML models into production is known as MLOps. In other words, MLOps standardizes processes whereas AIOps automates machines.MLOps standardizes processes whereas AIOps automates machines.AIOps is defined as the combination of big data and machine learning that automates IT operations activities including event correlation, outlier detection, and causality determination, according to Gartner, the company that first c