Worked on targeted AI improvements for the yairm210/OpenRA repository, focusing on refining unit behavior during combat scenarios. Addressed a critical bug by implementing logic in C# that enables units to disengage from targets they cannot attack, enhancing both action economy and gameplay realism. The solution involved adding armament suitability checks to the AI’s target selection process, ensuring units no longer pursue unattainable targets and reducing wasted actions. This update improved the tactical decision-making of unit AI and contributed to overall gameplay stability. Demonstrated skills in AI behavior programming, game development, and in-repo debugging within the constraints of the project.