Shay Friedman & Alex Suprun
CTO, CodeValue
About
Moaid is a Software Consultant, Code Jedi and an OzCode Evangelist at CodeValue. He views Software development as both an Art form and a Profession, and he is an advocate for Software Craftsmanship and Clean Code methodologies. Since the first time he had to debug a program, Moaid has aspired to master the art of debugging and to decrease debugging time as much as possible. In his blog (http://moaid.codes) Moaid writes about software, programming languages, software development tools and anything else he finds interesting.
About Shay
Shay, the CTO of CodeValue, is a veteran software architect with over 20 years of experience specializing in software architecture, technical leadership, and AI. He is the author of several technical books, including "IronRuby Unleashed," and is a frequent speaker at international technology conferences. Shay leads CodeValue’s technological vision, focusing on advancing development methodologies and ensuring the company remains at the forefront of transformative technologies like Enterprise AI.
About Alex
With over 15 years of technology experience, Alex Suprun brings a wealth of knowledge in backend development, software management, and software architecture to Architecture Next. He is renowned for his unique ability to combine big-picture strategic thinking with deep technical expertise, enabling him to drive innovative solutions in backend technology. In his session, Alex will share key insights drawing from his extensive background in end-to-end application development, offering attendees a comprehensive view of modern backend challenges and innovative solutions.
Sessions
Inside Claude Code: What Anthropic’s Own Codebase Teaches Us About Building AI Agents
Claude Code ships as a deceptively simple CLI - but its codebase reveals deliberate design choices that challenge how most teams build AI applications today. In this 45-minute session, we’ll walk through the patterns that actually matter: how the agent loop is structured, how tools are designed and exposed to the model, how context and state are managed without elaborate RAG pipelines, and why the team consistently chose minimal abstractions over frameworks. We’ll extract concrete lessons you can apply to your own AI tools and agentic systems - what to imitate, what to question, and what to skip.