
Past Event • Workshop
Safety and Guardrails in AI Coding Workshop, Build Club Brisbane
25 February 2026 • The Precinct, Fortitude Valley, Brisbane • Hosted by Build Club Brisbane
Building with AI requires more than speed. It requires structure. That message anchored Founder Syed Mosawi's session at Build Club Brisbane's Safety and Guardrails in AI Coding Workshop, held at The Precinct on 25 February 2026.
The workshop was organised by Jax Koh of Build Club, with sponsorship support from TEN13, An Vo and Sophie Robertson. The audience of developers and founders worked through how AI coding assistants now introduce a distinct class of risk, including hallucinated dependencies, security vulnerabilities and subtle logic flaws that move quickly into production when speed becomes the primary metric.
Where modern AI assisted development now sits
The session covered the intersection where building with AI now operates:
- Full stack thinking from front end through to backend, databases, permissions and money flows
- Intellectual property strategy including trade secrets and prompt design treated as protectable assets
- Privacy by design aligned with the Privacy Act 1988 and the General Data Protection Regulation (GDPR)
- Ongoing model testing, performance monitoring and meaningful human oversight
- Technical guardrails such as rate limiting, role based access control and strict input validation
- Terminal level Human in the Loop controls and sandboxed development environments that protect production systems from AI experimentation
Vibe coding as a discipline, not a vibe
The closing point of the session: vibe coding has graduated. What began as a way to ship faster is now a discipline that must operate within governance frameworks, security principles and IP strategy from day one. Treating AI assisted development as informal or experimental is no longer a defensible position, particularly for teams building anything that touches customer data, money or regulated workflows.
The takeaway for attendees was practical. Architecture, security, IP strategy and AI governance should be designed together as one stack rather than addressed separately after the system is already running.
From the workshop

