avatarBruno Aziza

Summary

The website content provides insights into the implementation and success of Data Mesh through MLOps, featuring case studies, expert opinions, and upcoming events.

Abstract

The article discusses the significance of MLOps in the success of AI projects, emphasizing the importance of a Data Mesh approach. It highlights a piece by George Mathew and the Insight Partners' team, which delves into the role of MLOps. The article also addresses the ongoing debate around the Data Mesh concept, refuting concerns about reinforcing data silos and emphasizing the need for a decentralized ownership model, distributed data development, and federated governance. Wayne Eckerson's perspective on rethinking the Data Mesh is presented, acknowledging its broader scope beyond traditional data virtualization tools. The article further promotes an upcoming conference where strategies for making the Data Mesh real will be shared. Success stories, such as Delivery Hero's replacement of over 20 data warehouses with a global Data Mesh using BigQuery, demonstrate the practical benefits of this approach at scale. Additionally, MoneySuperMarket's efficient data platform migration and utilization of cloud services like BigQuery, Dataflow, and PubSub are showcased as examples of effective data management.

Opinions

  • The author values the contribution of MLOps to AI project success, suggesting it as a key factor.
  • There is an acknowledgment that terminology should not overshadow the advancement of the field, particularly in the context of the Data Mesh debate.
  • The article positively references the idea that the Data Mesh is more than a technical solution; it is an evolving operating model that requires continuous adaptation.
  • The author agrees with Wayne Eckerson's view that the Data Mesh does not reinforce silos and instead focuses on the importance of federated governance and decentralized ownership.
  • The piece by Thinh Ha is cited to support the notion that treating Data Mesh as a fixed target rather than an evolving model leads to failure.
  • The author poses questions about the complexity of federation, the lack of culture, maturity, and talent, indicating these as challenges in the successful adoption of the Data Mesh.
  • The author is optimistic about the State of Data Mesh Conference as a platform for sharing practical strategies for Data Mesh implementation.
  • Delivery Hero's case is presented as a compelling example of the benefits of adopting a Data Mesh, highlighting the scalability and efficiency gains.
  • MoneySuperMarket's data platform transformation is highlighted to illustrate the impact of cloud services on streamlining data processes and saving costs.

How To Make The Data Mesh Real…

MLOps and how it can help AI succeed | Data Mesh Debates vs. Reality

If you’re new to this publication, this blog is YOUR Data, AI & Analytics Weekly Digest. I review the most popular data stories of the week & filter for you what’s HOT and what’s NOT.

If you like this, please consider commenting, liking and subscribing here.

1 — How MLOps can help AI project succeed

A good piece by my friend George Mathew and the great Insight Partners’ team, Sophie Beshar, Lonne Jaffe and Ganesh Bell. More here.

2 — Rethinking the Data Mesh

Great piece by Wayne Eckerson on the “Rethinking the DataMesh”. A few observations that I think are worth noting:

  1. While Wayne acknowledged the fact the “term data mesh conjures up a déjà vu of past distributed technologies”, he highlights the fact that focusing on terminology, or credit for the idea is not going to advance the field. Kudos for that!
  2. Wayne recognizes that the “Data Mesh” is “broader than a traditional data virtualization tool. It embeds the technology within a decentralized ownership model, distributed data development, and federated governance” YES! As Thinh Ha highlights it in his piece “10 reasons why you are not ready to adopt data mesh”, if you “ treat Data Mesh as a technical solution with a fixed target rather than an operating model that continuously evolves over time”…you will fail!
  3. Wayne poses at least 3 interesting questions in his post: does the Data Mesh reinforces silos?! (I don’t think it does in 2022), how do we deal with the complexity of federation and finally the issues that come with lack of culture, maturity and talent to succeed with the data mesh (I think that’s true of most technologies).

More here.

3 — How To Make The Data Mesh Real

Still Curious?! I’ll be presenting on How To Make The Data Mesh Real at the upcoming State of Data Mesh Conference. Register for free here.

4 — Success with the Data Mesh: Delivery Hero

20+ datawarehouses replaced with single global datamesh with BigQuery and unleashing the power of data with machine learning. With 791M orders in the third quarter of 2021 alone and revenue projected to reach $312,153 million in 2022. That’s what I call data delivery at scale! Story here, Video below (More here)

5 — Customer of the Week: MoneySuperMarket

2M events per day, 7M files, 150TB and a process that has gone from 11 steps to just 3–4 steps, and migrated from a different cloud in one weekend! Moneysupermarket Group’s Chief Data Architect Matthew Cresswell tells us how their customer data platform and the use of BigQuery, Dataflow, PubSub can help them save their customers money. More here, Video Below!

Business
Data
Analytics
Machine Learning
Artificial Intelligence
Recommended from ReadMedium
avatarFeruz Urazaliev
Effective Data Governance: Best Practices for Data Engineers

Data governance is crucial for ensuring data accuracy, security, and compliance within an organization. For data engineers, implementing…

7 min read
avatarShri Salem
Measuring the value of data management

A framework of now, next and near

13 min read
avatarAndriy Zabavskyy
Data Product in a Nutshell

A big-picture layout shaped by TOGAF’s domains

4 min read
avatarMikkel Dengsøe
How data observability fits into the different stages in the data pipeline

9 min read
avatarSagar Lad
Domain Driven Data Ownership and Architecture : Data Mesh Approach

Data Mesh architecture is a decentralised approach of creating data architecture which enables domain teams to manage their own domain…

4 min read
avatarChristopher Tao
Do Not Use LLM or Generative AI For These Use Cases

Choose correct AI techniques for the right use case families

7 min read