I just finished reading this book and decided to share my experience with it.
About the Book
Continuous Architecture in Practice is a sequel to Continuous Architecture both written by Murat Erder, Pierre Pureur, and Eion Woods. The authors attempt to address feedback from their 1st book by navigating the reader through the Trade Financs eXchange (TFX) case study.
The book starts with a brief introduction to the core concepts of Continuous Architecture, including its 6 principles, as well as its essential activities:
- Focus on quality attributes
- Drive architectural decisions
- Know your technical debt
- Implement feedback loops
From then on the book switches focus to common architectural concerns:
- Data
- Security
- Scalability
- Performance
- Resilience
- Emerging Technologies
Each concern is tackled in a separate chapter that features an introductory quote, a definition along with some historical context, a list of issues an someone should keep in mind when architecting for that concern, a list of tactics, and a further reading section.
My Opinion
Overall, I liked reading the book, it gave me a lot of inspiration and a desire to learn more about particular topics. The book assumes a certain level of familiarity with software engineering which helps it focus on general concerns and avoid implemention specific details.
I particularly liked the Emerging Technlogies chapter as it offers a healthy view on AI, ML, and shared ledgers. It helps remove the fairy dust that's blinding our industry, and instead focuses on meaningful changes that actually provide value to a product.
Also, as someone with experience in software security I appreciated the focus on shifting left security concerns.
Who Should Read This
In my opinion, every software engineer can benefit from reading this at some point in their career. However, I wouldn't recommend it to a junior, if you are not already familiar with the topics covered in the book it could be intemediating.
PS I am not affiliated with the book or its authors in any way. I am just a person that read a book they liked wanted to share my experience.
I know you are asking for something different, but since there are already a few good answers, allow me to instead to reject the premise and give you a different.
It's not impossible to implement an AI solution within the context your provided. The problem is that it's going to be expensive. However, you can offer to deliver something smaller, focus on the smallest but valuable contribution you can make. While cleaning up the data is still going to be a hell of task, if the scope is small enough it can be achievable. Then, you can communicate the difficulty to scale due to data issues which can help management undestand the importance of prioritizing data quality.
If you have a bunch of sales data, maybe you can focus on deriving purchase patterns and build a simple recommendations engine. If you want to focus on marketing, you could try lead classification. Ideas depend on the domain of the company you work for.