Gus Shaw Stewart

Gus Shaw Stewart

Strategic Solutions Engineer Port.io
Gus Shaw Stewart

Gus Shaw Stewart is a Strategic Solutions Engineer at Port who spends his days convincing platform teams that AI agents don't have to be terrifying. Before Port, Gus was at GitHub helping enterprise teams tame Copilot in high-security environments. He's the founder of London's Platform Nation community, a frequent conference speaker, and a former full-stack developer who still codes side projects, paints oil portraits, and pretends he's going to retire from amateur football.

CONFERENCE DAY ONE - Wednesday 21st October

4:00 PM Afternoon Keynote Presentation: Software Factories: Scaling Agentic Delivery

As AI agents move from assistants to primary builders, teams face a new governance problem: how do you scale agentic workflows across hundreds of repos, multiple teams, and different tech stacks? The single-repository playbook breaks down. This session covers what enterprises actually need - orchestration, context management, cost control, and audit trails - to turn agentic delivery into a reliable, governed platform capability. We'll walk through the architecture of a software factory and show how platform teams can operationalize AI-native delivery without losing control.

The Shift
  • Code is no longer the bottleneck; planning, review, and governance are
  • Traditional SDLC workflows built around coding time need redesign
  • Agentic workflows flatten hierarchy- agents route work autonomously
What Breaks at Scale
  • Single-repo playbooks don't translate to multi-team, multi-repo environments
  • Context fragmentation (scattered docs, tribal knowledge, tool silos)
  • No visibility: cost per feature, agent behavior, artifact quality, audit trails
  • Governance defined at repo level can't enforce org-wide policy
  • Agent access control and permissions become critical
Architecture of an Enterprise Software Factory
  • Workflow orchestration (intent > spec > plan > build > test > deploy as automated pipeline)
  • Context layer (centralized knowledge graph vs. scattered markdown files)
  • Agent registry and governance (who can use what, with what constraints)
  • Observability tier (cost tracking, quality scoring, lead time, failure modes)
  • Human-in-the-loop decision points (where humans actually add value)
Live Demo
  • Feature request moves through intent > spec > plan workflow
  • Control plane shows org-level visibility across all teams, all factories
For Platform Teams
  • This is your next evolution - you now own agentic delivery governance
  • Start small: one workflow, one team, get the pattern right
  • Build incrementally: add repos, add teams, add policies as you scale

Check out the incredible speaker line-up to see who will be joining Gus.

Download The Latest Agenda