Roberto is the Manager of the Development Core team at MetService. A practitioner-leader based in Wellington, he focuses on AI-first modernisation of live software systems, where constraints, risk, and institutional trust matter. His work centres on integrating AI into software engineering in ways that preserve ownership, clarity, and architectural control. He is the originator of frameworks such as Micromanaged-Driven Development (MMDD) and MDSR (Moderniser), which explore how GenAI can support reasoning, sequencing, and decision-making rather than replace human judgment. Over more than two decades, Roberto has worked across multiple domains, including Python development — contributing to the Plone Foundation — mobile development, and cloud-based systems. He has spoken at international conferences including PyCon US, as well as events across South America, Europe, and New Zealand, including AWS Community Day Wellington, AWS Public Sector Day (2025), and Mobile Refresh.
GenAI is already being used in everyday coding workflows — but many teams are adopting it without a clear mental model for when it helps, when it hurts, and who remains responsible.
This talk is for software developers and technical leaders who are using — or being
asked to use — GenAI in day-to-day coding, design, and review work. Rather than treating GenAI as a junior engineer or autonomous system, the session frames it as a reasoning tool whose value depends on how humans constrain, interpret, and remain accountable for its output.
With the right framing, GenAI can accelerate learning, exploration, and problem-solving. With the wrong one, it can amplify misunderstanding, create false confidence, and quietly erode architectural control. This talk explores where GenAI provides genuine leverage in coding work, where it introduces new risks, and why familiar delegation and automation models often break down.
Core themes
- GenAI as a reasoning tool rather than a replacement developer
- Where GenAI helps in coding and design — and where it creates risk
- Why responsibility and ownership become blurred in AI-assisted workflows
- Human-in-the-loop patterns and deliberate sequencing
What attendees will take away
- Practical rules of thumb for when to use GenAI in coding and design work — and when not to
- Patterns for keeping humans accountable in AI-assisted development
- Common failure modes to avoid when introducing GenAI into software teams
