Govern autonomous engineering inside your Azure tenant | ODSP2201
Sean Patten presents how BuildEngine by Eido fits autonomous engineering into an existing Microsoft environment, with execution and control kept inside the organization’s Azure tenant.
Governed engineering workflows
This Microsoft Ignite 2026 session focuses on the practical concerns of adopting AI-driven engineering: cost, security, and auditability. It covers parallel engineering workflows with human approval gates and audit history, so teams can retain oversight as agents carry out development tasks.
The Azure-based architecture uses isolated agent environments. Windows 365 for Agents and existing Intune policies provide the environment and policy context for agent work, alongside model choice and cost controls.
Model selection through Microsoft Foundry
Microsoft Foundry provides access to models from OpenAI, Claude, DeepSeek, and other providers. The session covers using different models for different workflow stages rather than relying on a single model throughout the engineering process.
The demonstration begins with deploying DeepSeek and Astra models in Azure Foundry, then moves to creating and configuring an AI-powered development workflow.
Demonstration: from feature request to QA
Patten follows a feature request through the engineering workflow:
- Automated design and refinement.
- Build and testing.
- Code review.
- Quality assurance using a Windows 365 Cloud PC.
The focus is on combining automated execution with governance: isolated environments, human approvals, auditability, and control within the Azure tenant.
Session chapters
- 00:00 — Introduction to autonomous engineering in Azure with Eido Build.
- 01:22 — AI engineering challenges: cost, security, and auditability.
- 02:48 — Parallel AI-driven engineering workflows with human approvals.
- 04:01 — Azure-based architecture, isolated agent environments, and governance.
- 05:29 — Multiple AI models through Microsoft Foundry for different workflow stages.
- 07:26 — Demonstration: deploying DeepSeek and Astra models in Azure Foundry.
- 08:36 — Creating and configuring an AI-powered development workflow.
- 12:41 — Running a feature request through automated design and refinement.
- 16:49 — Automated build, testing, code review, and QA with Windows 365 Cloud PC.
- 21:18 — Key lessons on governed autonomous engineering.
The session is a pre-recorded, intermediate-level Cloud & AI presentation, delivered in English (US), with session code ODSP2201.