Software Architecture Superstream: Architecture Patterns and Antipatterns for AI
Published by O'Reilly Media, Inc.
Exploring the Foundations of AI-Augmented Architecture
Understanding how to integrate generative AI technologies into application architecture is now essential to your success as a software architect, but figuring out how, where, and when to introduce them into your solutions can be daunting. We’ve got you covered.
Join host Neal Ford and a slate of experts to explore the combination of techniques, tools, and mechanisms that will help you introduce generative AI-powered features and capabilities into your architecture. You’ll learn common patterns, antipatterns, strategies, use cases, and more that you can employ immediately to harness the power of generative AI.
What you’ll learn and how you can apply it
- Design modern systems that work well with generative and agent-based AI
- Understand where AI tools can add real value in software architecture—and where they can’t
- Learn what to consider when using vector databases and other AI-era infrastructure
Recommended follow-up
- Read Building Applications with AI Agents (early release book)
- Read Building Micro-Frontends, second edition (early release book)
- Take Building AI Agents with Model Context Protocol (MCP) (live online course with Lucas Soares)
Schedule
The time frames are only estimates and may vary according to how the class is progressing.
Introduction – Neal Ford (5 minutes)
Neal Ford welcomes you to the Software Architecture Superstream.
Emerging Patterns in Agentic AI – Bharani Subramaniam (35 minutes)
Join Bharani Subramaniam, CTO for India and the Middle East at Thoughtworks, to explore the evolving landscape of agentic AI systems, including key enabling technologies like the Model Context Protocol, the role of action spaces—particularly code generation—and the impact of tool integration. He also covers memory architectures, from short-term recall to long-term persistence, as essential components of learning and context management. Concluding the session, he offers a forward-looking view of emerging model architectures.
Beyond the Hype: AI Tools in Micro-Frontend Architectures – Luca Mezzalira (35 minutes)
AI won’t replace developers but it’s changing how we work. Luca Mezzalira, principal serverless specialist solutions architect at AWS, explores where today’s AI tools actually help in a micro-frontend architecture—from prototyping and testing to migration and automation. He’ll cover real use cases, common pitfalls, and a playbook for getting useful results without giving up architectural control.
Break (5 minutes)
Mind Your Language Models: An Approach to Architecting Intelligent Systems – Nischal HP (35 minutes)
As large language models move from proof of concept to production, the need for robust architectural strategies rises. AI and technology leader Nischal HP offers practical guidance on designing and deploying intelligent systems using LLMs, based on real-world experience. Topics include fine-tuning, zero-shot learning, prompt engineering, and managing common integration challenges such as latency, tokenization, and safety.
Architecting for Vectors: Evolving Database Design for the AI Era – Dikshant Adhikari and David Bressler (Sponsored by Cockroach Labs) (30 minutes)
As AI becomes core to modern applications, architects face mounting pressure to support vector search without compromising transactional guarantees. These demands are pushing legacy database architectures to their limits and prompting a reassessment of how to build for both scale and consistency. David Bressler and Dikshant Adhikari explore the trade-offs between specialized vector stores and integrated approaches, emerging patterns in distributed systems, and how CockroachDB is addressing these challenges.
This session will be followed by a 30-minute Q&A in a breakout room. Stop by if you have more questions for Dikshant and David.
Break (5 minutes)
Why Build Event-Driven AI Systems? – Adi Polak (35 minutes)
Understand what an AI agent is, what a modern AI system looks like, and how to architect that system for production using event-driven microservices. Adi Polak, author and official Databricks ambassador, explores agentic architecture in multi-agent AI systems, emphasizing the roles of controller, planner, and judge for intelligent coordination. She’ll also discuss building scalable, event-driven microservices with Kafka and Flink and highlight best practices for RAG pipelines, governance, and observability.
Why MCP Really Is a Big Deal – Tim Berglund (35 minutes)
Tim Berglund, VP of developer relations at Confluent, is back at the lightboard with MCP (Model Context Protocol). MCP is a big deal, but most people are missing the point. It's not just about enhancing desktop applications with agentic functionality—it's about exposing reusable tools and resources (including real-time data from sources like Apache Kafka) to build agentic microservices. Tim describes how to go beyond basic LLM chatbots to dynamic, problem-solving systems that deliver real value in real professional settings.
Closing Remarks – Neal Ford (5 minutes)
Neal Ford closes out today’s event.
Your Hosts and Selected Speakers
Neal Ford
Neal Ford is an independent software consultant, best-selling author, speaker, and innovator in building concrete verification mechanisms for software architecture, for humans or agents. He’s an internationally recognized expert on software development and delivery, especially at the intersection of agile engineering techniques and software architecture. Neal has authored magazine articles, nine books (and counting), and dozens of video presentations, and has spoken at hundreds of developers’ conferences worldwide. His topics include software architecture, continuous delivery, functional programming, cutting-edge software engineering, and improving technical presentations. Check out his web site at Nealford.com.
Bharani Subramaniam
Bharani Subramaniam is CTO for India and the Middle East at Thoughtworks, where he shapes new pursuits and builds tech strategies with clients, contributing to the Thoughtworks Technology Radar. He enjoys building low latency, high throughput applications and distributed systems, and he’s passionate about simplifying system designs, data engineering, and real-time analytics.
Luca Mezzalira
Luca Mezzalira is a software architect, author, and technology leader with more than 20 years of experience helping engineering organizations design distributed systems that scale, both technically and organizationally. He has held architecture leadership roles at DAZN and AWS, where he helped organizations modernize software delivery through distributed systems, serverless architectures, and cloud native technologies.
Luca is widely recognized for his pioneering work on micro-frontends and is the author of the O'Reilly book Building Micro-Frontends, a foundational resource for engineering teams modernizing their frontend architectures. Today, his research focuses on how AI is reshaping software architecture, engineering workflows, and the evolving role of architects in an AI-native world.
A frequent international speaker, Luca has delivered more than 350 talks and workshops at events including AWS re:Invent, QCon, GOTO, NDC, and O'Reilly Superstream. He is also the creator of the Dear Architects, Token by Token, and Micro-Frontends newsletters, and hosts the Dear Architects and Micro-Frontends in the Trenches podcasts.
Nischal HP
Nischal is an AI and technology leader with 15 years of experience translating complex AI research into high-impact products. He specializes in building and leading high-performance engineering organizations and leading teams to craft sophisticated AI solutions, including designing the MLOps platforms and robust cloud infrastructure necessary to productize cutting-edge models and to harness the power of large language models for novel applications.
Dikshant Adhikari
Dikshant Adhikari is a senior product manager at Cockroach Labs, where he has overseen developer tools, database schema, roles/privileges, and PostgreSQL compatibility. His current focus is CockroachDB’s vector capabilities. Previously, Dikshant worked at DigitalOcean and IBM as a software engineer.
David Bressler
David Bressler is a staff product marketing manager at Cockroach Labs. He has worked in 26 countries and is an accomplished public speaker. He graduated with distinction with an MBA from NYU and is also an athlete, expedition cave and shipwreck diver, and a two-time medalist in the 1989 World Maccabiah Games in Israel.
Adi Polak
Adi Polak is an official Databricks ambassador, author of the book Scaling Machine Learning with Spark, and a respected worldwide presenter. As a data practitioner, she developed algorithms to solve real-world problems using machine-learning techniques. As an engineer, she brought her hands-on ML experience into various Fortune 500 companies' products and services by building upon cutting-edge and emerging technologies.
Tim Berglund
Tim Berglund is VP of developer relations at Confluent, where he and his team work to make streaming data and Confluent’s emerging tool set accessible to all developers. He’s a regular speaker at conferences and a presence on YouTube, explaining complex technology topics in a user-friendly way. He lives with his wife in California. He has three grown children, three step-children, and five grandchildren.
