Uploaded August 2026 | Updated September 2026, 2 weeks ago
When volume of code is no longer a bottleneck, it’s the constraints you choose that determine whether an engineering team is successful.
Microsoft’s Maxim Salnikov has helped hundreds of companies transform the way they work with AI and often speaks about its impact on engineering.
As he watches companies experiment with these new technologies, Maxim has seen both positive patterns – teams using AI well – and what he calls “anti-patterns”. These are when teams fall into traps that set them back needlessly, like siloing AI expertise in a small group of specialists or qualifying AI usage based on code importance.
In this session, Maxim joins JetBrains’s Ciara Byrne to discuss changing developer roles in the age of agentic AI and best practices for teams looking to adopt it responsibly.
Topics covered include:
👩💻 How developer roles are changing with each new generation of agents
🏗️ The architecture teams need to make agents successful
🔍 How to address ballooning review cycles with constraints engineering
🎯 Context engineering for impact instead of output
When volume of code is no longer a bottleneck, it’s the constraints you choose that determine whether an engineering team is successful.
Microsoft’s Maxim Salnikov has helped hundreds of companies transform the way they work with AI and often speaks about its impact on engineering.
As he watches companies experiment with these new technologies, Maxim has seen both positive patterns – teams using AI well – and what he calls “anti-patterns”. These are when teams fall into traps that set them back needlessly, like siloing AI expertise in a small group of specialists or qualifying AI usage based on code importance.
In this session, Maxim joins JetBrains’s Ciara Byrne to discuss changing developer roles in the age of agentic AI and best practices for teams looking to adopt it responsibly.
Topics covered include:
👩💻 How developer roles are changing with each new generation of agents
🏗️ The architecture teams need to make agents successful
🔍 How to address ballooning review cycles with constraints engineering
🎯 Context engineering for impact instead of output
