Building aIDP

What You'll Learn

• ›What an automated software factory is, and which parts of the development flow can run end to end without a human driving them
• ›How to build work intake from multiple sources: code quality findings from SonarQube, the product backlog, and other queues
• ›How to orchestrate several AI colleagues at once: assigning work, running them in parallel, and keeping them out of each other's way
• ›How to design the pipeline from an incoming work item to a merged, deployed change
• ›How to build a dashboard that shows throughput, success rate, and where the factory is stuck
• ›Where to keep humans in the loop, and how to set the limits the factory operates within

How It Works

Participants build a working aIDP during the day. We wire up work intake from a code quality tool and a product backlog, orchestrate a set of AI colleagues to pick up and complete the items, and build the dashboard that shows what the factory is producing and how well it is doing it.

• ›What aIDP is: from individual agents to an automated software factory
• ›Work intake: SonarQube findings, product backlog, and other sources
• ›Orchestrating AI colleagues: assignment, parallelism, and conflict handling
• ›The pipeline: from work item to merged and deployed change
• ›The dashboard: throughput, success rate, and bottlenecks
• ›Human oversight: approval gates, limits, and escalation

Outcome

By the end of the training, participants will: Understand how an automated software factory is put together, Have built a working aIDP that takes in work from multiple sources and completes it, Know how to orchestrate multiple AI colleagues without them getting in each other's way, Be able to measure the factory's performance and spot where it stalls