ICS - Integrated Computer Solutions, Inc.

AI coding tools are good at writing code in general, but not so good at writing code the way your team writes it. In our latest post, part 1 of a 3-part series called How To Make AI Code Your Way, we explain how we used project-specific Claude Code skills to teach an AI agent our Qt/QML architecture, dependency injection patterns and testing conventions for embedded Linux applications. A small amount of structure made a big difference in code consistency. Read: Teach Claude Code Your Project-Specific Skills.

In our next webinar, we’ll explore how memory is becoming the core design challenge for agentic AI—from context limits and vector search to graph-based models and external memory systems. More importantly, we’ll tackle a critical question: what should AI forget? 🤔

Join us to learn how leading approaches balance knowledge, performance, and privacy—and what’s next for building truly intelligent systems. https://bit.ly/4ceFmXh