Sign In
Register

Partnership opportunities

Request your invite

Call to action
Your text goes here. Insert your content, thoughts, or information in this space.
Button

Back to speakers

Gugulethu
Ncube
Senior Director, Engineering
Delivery Hero
Gugulethu Ncube is Senior Director of Engineering at Delivery Hero, where he leads engineering teams responsible for building and scaling technology that supports one of the world's largest local commerce platforms. He has extensive experience leading distributed engineering organizations, driving technical strategy, and delivering software that creates measurable business impact. A former startup Co-Founder and CTO, Gugulethu has built and led engineering teams across multiple regions, establishing agile ways of working, scalable engineering practices, and high-performing cultures. He is a strong advocate for empowering teams through autonomy, clear ownership, accountability, and transparency, enabling engineers to move quickly while maintaining operational excellence. Passionate about technology and software engineering, Gugulethu continues to code in his spare time and regularly shares insights on engineering, leadership, and software development through his technical blog.
Button
15 September 2026 15:30 - 16:00
Managing Intelligence, cost and reliability in Agentic AI
Agentic AI systems are moving into production, where capability alone is not enough. Teams also need to decide what counts as a successful outcome, where more intelligence is worth the cost, and how much time, authority and autonomy an agent should have. This talk looks at practical ways to run agentic systems with clearer operational boundaries. It covers outcome-based measurement, dynamic use of models and reasoning, execution limits, verification, human handover and cost control. A key theme is that a completed run is not necessarily a completed job, and that model cost alone can be a poor measure of system economics. The focus is on the engineering and operating decisions behind reliable agentic systems. It also looks at how these decisions connect technical performance with business outcomes, especially when failure, delay or unnecessary human intervention has a real cost.